ELECTROTECHNICA & ELECTRONICA E+E...2018/01/02 · ELECTROTECHNICA & ELECTRONICA E+E Vol. 53. No...

ELECTROTECHNICA & ELECTRONICA E+E Vol. 53. No 1-2/2018 Monthly scientific and technical journal Published by: The Union of Electronics, Electrical Engineering and Telecommunications /CEEC/, BULGARIA

Editor-in-chief: Prof. Ivan Yatchev, Bulgaria

Deputy Editor-in-chief: Prof. Seferin Mirtchev, Bulgaria Editorial Board: Prof. Anatoliy Aleksandrov, Bulgaria Acad. Prof. Chavdar Rumenin, Bulgaria Prof. Christian Magele, Austria Prof. Georgi Stoyanov, Bulgaria Assoc. Prof. Evdokia Sotirova, Bulgaria Prof. Ewen Ritchie, Denmark Prof. Hannes Toepfer, Germany Dr. Hartmut Brauer, Germany Prof. Marin Hristov, Bulgaria Prof. Maurizio Repetto, Italy Prof. Mihail Antchev, Bulgaria Prof. Nikolay Mihailov, Bulgaria Prof. Radi Romansky, Bulgaria Prof. Rosen Vasilev, Bulgaria Prof. Takeshi Tanaka, Japan Prof. Ventsislav Valchev, Bulgaria Dr. Vladimir Shelyagin, Ukraine Acad. Prof. Yuriy I. Yakymenko, Ukraine Assoc. Prof. Zahari Zarkov, Bulgaria Advisory Board: Prof. Dimitar Rachev, Bulgaria Prof. Emil Sokolov, Bulgaria Corr. Member Prof. Georgi Mladenov, Bulgaria Prof. Ivan Dotsinski, Bulgaria Assoc. Prof. Ivan Vassilev, Bulgaria Assoc. Prof. Ivan Shishkov, Bulgaria Prof. Jecho Kostov, Bulgaria Prof. Lyudmil Dakovski, Bulgaria Prof. Mintcho Mintchev, Bulgaria Prof. Nickolay Velchev, Bulgaria Assoc. Prof. Petar Popov, Bulgaria Prof. Rumena Stancheva, Bulgaria Prof. Stefan Tabakov, Bulgaria Technical editor: Zahari Zarkov Corresponding address: 108 Rakovski St. Sofia 1000 BULGARIA Tel. +359 2 987 97 67 e-mail: [email protected] http://epluse.fnts.bg ISSN 0861-4717

C O N T E N T S

ELECTRONICS

Marin Marinov Smart sensor node for on-line particulate matter monitoring 1

Kiril Ivanov, Ivo Iliev, Serafim Tabakov Plasma based sterilization using new multipoint ignition accessory 10

Valeri Petkov, Nikolay Madzharov Design of axial rotary wireless energy transmitter 16

Viktor Tomov, Ivo Iliev, Vessela Krasteva Current waveform analysis for arc control in electrosurgical devices 22

Galia Marinova, Zdravka Tchobanova Study of the factors influencing power consumption of FPGA-based designs 28

TELECOMMUNICATIONS SCIENCE

Evelina Pencheva, Denitsa Kireva, Ivaylo Atanasov Multi-access edge service for application-initiated offload 35

INNOVATIVE TECHNOLOGIES

Galia Nedeltcheva, Elena Shoikova Innovation through Design Thinking, User Experience and Agile: Towards Cooperation Framework 42

In 2018, the journal Electrotechnica & Electronica is published with the financial support of the Scientific Research Fund at the Ministry of Education and Science.

ELECTRONICS

Smart sensor node for on-line particulate matter monitoring

Marin Marinov

Increased concentration of particulate matter in big cities and agglomerations is one of the main

risk factors for the health of the residents. Conventional systems for particulate matter (PM) monitoring have significant limitations, especially with respect to the costs of their installation and maintenance. The number of these monitor stations is limited but concentration of PM changes fast, non-linearly, within wide limits and is determined by multiple factors. In this paper, we propose an approach based on cost-effective sensor nodes for on-line measuring and obtaining information about PM concentration. The acquired data can be analyzed to evaluate the local distribution of pollutant density and can be shared through web platforms that use standard protocols. The adoption of such an approach allows for a detailed study of the levels of pollution and its sources.

Keywords – Air quality directive, Particulate matter monitoring, Sensor system, Raspberry Pi, Optical particle counter

Мобилна система за мониторинг на фини прахови частици (Марин Б. Маринов). Повишената концентрация на фини прахови частици в големите градове и селищни агломерации е един от основните рискови фактори за здравето на хората. Конвенционалните системи за мониторинг на параметрите на въздуха имат значителни ограничения, особено по отношение на разходите за тяхната инсталация и поддръжка. Броят на тези станции е ограничен, а концентрацията на ФПЧ се изменя бързо, нелинейно, в широки граници и зависи от множество фактор В тази статия, ние предлагаме подход, базирана на нискобюджетни сензори, за измерване в онлайн и получаване на детайлна информация за концентрация на ФПЧ. Получените данни могат да бъдат анализирани, за да се оцени локалното разпределение на плътността на замърсителите и да се споделят актуални стойности чрез уеб платформи, използващи стандартни протоколи. Използването на този подход дава възможност за по-детайлно изследване на нивата на замърсяване и на източниците му.

1. Introduction and motivation

Aerosols are two-phase systems, consisting of suspended solid or liquid phase, and surrounding gas phase. They are formed through the conversion of gases into particles or through the disintegration of liquids or solids into finer constituents.

The need to measure basic aerosol parameters has increased dramatically in the last decade. This is mainly due to their harmful effect on our environment and on public health.

Legislation requires that particle emissions and workplace particle concentrations are measured to ensure that the set limits are observed and the public is not exposed to undesirable aerosol concentrations [1].

Recently, the World Health Organization (WHO) estimated that exposure to particulate matter ( ) in indoor and outdoor air causes about 2 million deaths per year.

Although most highly developed countries have had considerable success in reducing air pollution through different policies, a significant part of Europeans still live in regions where regulated limits are constantly exceeded and a significant decline in life expectancy is observed [2, 3].

Increasingly complex and demanding regulations to mitigate pollution mean that aerosol measurements are becoming more and more time- and resource-intensive. These measurements often require more than basic knowledge of how to conduct and interpret the measurements. The elaboration of cost effective mitigation strategies depends on reliable and accurate measurement of the physical and chemical characteristics of aerosols.

High temporal and spatial resolution is mandatory for obtaining reliable data to be used for implementing

“Е+Е”, vol. 53, 1-2, 2018 1

policies and measures that would protect the health of the citizens.

The monitoring networks that are currently in use consist of a limited number of fixed stations.

The costs associated with the construction of an air quality monitoring station are about 200 000 Euro and those for its yearly maintenance amount to about 25 000 Euro [4]. As a result, the number of stations in most world metropolises is limited (Table 1). For instance, in Sofia there are only 10 stations covering an area of 1310km with a population of about 1.5million [5].

Table 1 Number of measurement stations and area coverage in some

metropolises [6], [7].

City # of stations Area, Beijing, China 18 ∼16 000

Berlin, Germany 12 ∼890

Mumbai, India 7 ∼440 New York City 13 ∼1 200 Sofia 10 ∼1 310

The density of control points is inadequate and the

data does not reflect personal health risk. Practically, there is a deficiency of trustworthy

information about the concentration of the measured parameters and how they vary over time even in the closest vicinity of the fixed stations and no individualized personal information can be obtained. Recent research has shown that there is a clear tendency for a shift in the current monitoring paradigm. Building up wireless networks with mobile and cost effective sensors is gaining popularity as an alternative way of monitoring air parameters with a much higher resolution capability. This approach makes it possible to collect personal-specific indicative information that enables people to monitor the quality of the air they breathe [6].

Due to the strict quality control exercised over monitoring networks it is possible to compare results obtained with different instruments and from different networks. By analogy, prior to using low cost sensors, their characteristics need also to be assessed. Information needs to be acquired about numerous parameters such as:

• concentration range and detection limit, • stability of response - comparability and

variability between individual sensor devices, • influence of temperature, humidity, atmospheric

pressure, • sensitivity to detect temporal variations.

It may be necessary to perform complex procedures

of calibration checks prior to deployment. The first step is to understand the capabilities of the sensor devices, and the characteristics of the information they provide [6].

These considerations necessitate the adoption of additional measures in order to acquire more detailed data about the levels of pollution and main emittents. Our study is part of the measures taken in this respect.

In this paper, we propose the acquisition of real-time and fine-grained information throughout a city, by means of mobile low-cost sensor nodes. The use of cost effective particulate matter monitoring systems allows for a detailed study of the levels of major pollutants and their sources.

2. Particulate matter monitoring – basic parameters and standards

At present, there is no uniform standard or technique of measuring particulate matter because aerosols and particles, which are the result of natural and man-made processes and are found in ambient and industrial environments exist in varied sizes, shapes, densities, chemical compositions, and biological properties.

2.1 Basic parameters

There are a number of relevant parameters of in ambient air that should be taken into account:

• particle number concentration, • total mass concentration of selected fractions of

particulate matter, • particle size distribution, • 24-hour variations of concentrations with peak

values, • chemical composition, etc.

Currently, the most common health relevant metric is mass related to particle size, which is represented as

. In most sources, the abbreviation is incorrectly defined as "all fine particles with size smaller than ". The correct definition is “particulate matter which passes through a size-sensitive inlet with a 50% efficiency at aerodynamic diameter” [7].

limits and targets for 24 hours and annual averages differ significantly from country to country. This is shown in Table 2 which provides examples of

standards and objectives of some countries and the World Health Organization.

Apart from diverse national and international standards, there is the challenge that air quality varies non-linearly by locations and a great many factors, such as the weather conditions, traffic, land use, etc., which affect it and make it very complicated to be modeled.

2 “Е+Е”, vol. 53, 1-2, 2018

In fact, unless there is a monitoring station, we do not really know what the air quality of a location is.

Table 2 Standards and objectives for monitoring in urban areas in

some counties [6].

Country

fraction Averaging period

Limit /Standard /

EU Annual 40 24 h 50 . Annual 25/20 US 24 h 150 . Annual 12 24 h 35 China Annual 70 24 h 150 . Annual 35 24 h 75 India Annual 60 24 h 100 . Annual 40 24 h 60 WHO Annual 20 24 h 50 . Annual 10 24 h 25

2.2. Fractions of particulate matter

The size of the particles is directly related to the potential they have to cause health problems - the finer the particles, the more difficult they are to disperse and the deeper they can penetrate into the lungs and even into the blood stream thus causing more harm. enters the respiratory tract, and has been associated with health risks such as bronchitis, asthma, and upper respiratory tract infections. aggravates symptoms of existing diseases rather than triggers new conditions.

Fine particles in the . size range get into the respiratory tract and can reach the lungs and the blood stream causing cardiovascular problems.

The statistics show that 7% of the urban population in the EU-28 was exposed to . levels above the EU limit value, and approximately 82% was exposed to concentrations exceeding the stricter WHO AQG value for . in 2015 [8, 9].

, which are so much smaller than . , can have stronger effects and give rise to lasting conditions, such as people predisposed to heart diseases. Studies in the west have shown that can lead to premature

births and affect prenatal development. Therefore, the measurement of is of great

interest for recent studies. Intensive work is being done to establish standards and targets for monitoring .

Advances in embedded systems and new sensors technologies have made it possible for a new generation of low-cost monitoring systems to emerge. Portable and autonomous sensors have the potential to take measurements with sufficient accuracy and in this way to capture effectively the spatial variability of the air pollutants. The number of those commercially available devices has increased considerably over the last five years although the quality of the data which they provide is still questionable [10]. The main goal of our study is to test the quality of the data obtained by semi-professional sensors and compare these results with the measurements of professional devices or those reported by the official authority stations. We want to find out whether such cost-effective systems can provide reliable results and indications about air quality and can be used in practice.

3. Design of the proposed monitoring system

The overall system architecture and the design of hardware and software components are presented in details in this section.

3.1. Design Requirements

• High reliability and availability of the device for long-term measurements.

• Ability to measure not only but also . and .

• Use of off-the-shelf cost-effective components for wireless sensor networks implementation and support of different communication standards (e.g., Zigbee, 802.15.4, WiFi, BLE) compliant with pervasive deployment;

• Availability of various integrated programming environments (MATLAB/LabVIEW) to reduce the implementation time.

• Computing power to perform on-board calculations, scalable architecture that supports easy expansions with peripherals e.g. sound cards, microphones, environmental parameter sensors.

• Capability for remote status monitoring, GPS localization and software updates.

• Long-term battery powered operation. Fig. 1 shows the overall system architecture of an

environmental monitoring network system that we have developed. The system consists of a Single Board Computer (Raspberry Pi 3), a GPS module with

“Е+Е”, vol. 53, 1-2, 2018 3

integrated RTC and an Optical Particle Counter OPC-N2 for measuring , . and .

Fig. 1. Block diagram of the developed node.

3.2. Hardware

The Raspberry Pi 3 Platform Taking into account the design requirements and the

needed functionalities a scalable architecture based on the Raspberry Pi 3 (RPi) platform is proposed.

It has the following characteristics: it consists of the Broadcom BCM2837 System on a Chip, including 4 cores ARM Cortex-A53, running on a 1.2 processor, a Graphic Processing Unit (GPU), 1 of RAM with a microSD memory card slot.

The Raspberry Pi platform has a number of advantages such as good computing power, high versatility and is enabled for MATLAB/LabVIEW programming environment. The low power consumption and its price allow for the design of numerous devices based on this platform. Those qualities, together with the continuous hardware upgrades, made the Raspberry Pi the selected option for the prototype development.

Raspberry Pi has various options for data transfer - mainly via Ethernet, Wi-Fi or Bluetooth connectivity. The most appropriate alternative is the use of the IEEE 802.11n-based Wi-Fi built-in chip. The module operates at 2.4 and provides up to 150 / transfer rates. The distance of the station to the access point can be up to 300 but this strongly depends on the terrain and surrounding background radio noise. In this particular case, tests with a direct visibility of up to 50 were conducted [11].

Thanks to the integrated Wi-Fi module, there is a choice from two options for connecting to the internet – in a wireless mode or via an Ethernet cable. Connection through Ethernet is usually more stable but requires access to a wired network. The tests of the system show that the Wi-Fi connection is a good choice when the signal strength is strong.

Sensor

The sensor is Alphasense OPC-N2 [2]. It is cost efficient and has good performance when compared with other units (Fig. 2). The sensor measures particle size and aerosol mass concentration by detecting the light scattered from individual particles. The sampled air is drawn through the laser beam with a wavelength of 658 .

Fig. 2. The Alphasense OPC-N2 particle counter.

Here instead of a pump, the instrument uses an electrical fan, the so called ‘pump-less’ design, which requires no replaceable pump-protection filter. The scattered light is picked up by an elliptical mirror and focused on the dual-element photodetector. Like most commercial Optical Particle Counters, all particles, regardless of their shape are assumed to be spherical and are therefore assigned a ‘spherical equivalent size’. The relationship between particle size and scattered intensity at any given scattering angle is complex and exactly described by Mie’s theory [12].

The sensor classifies each particle size, at rates of up to 10000 particles per second, and records particle size in one of 16 “bins” covering size range from 0.38 microns to 17 microns in diameter. This data is used to calculate the mass of airborne particles per unit volume of air, expressed in / .

4 “Е+Е”, vol. 53, 1-2, 2018

The OPC-N2 requires a 4.8 to 5.2 Volt DC supply with minimum electrical noise and at least 175 continuous current [13]. The connection with Raspberry Pi 3 is made via the SPI bus, using SPI communication protocol.

GPS module

The GPS module gives information about the position of the node and its altitude. The attached GPS receiver is based on u-blox NEO-7M, with high-gain active antenna. It provides maximum sensitivity (−161 ) while maintaining low system power (35 operating current). An important part of the module is the build-in real-time clock, which keeps track of the current time, while the primary source of power is off or unavailable. Transfer of the data between a microcomputer and the GPS module is achieved through serial UART communication.

Power Supply

The power source of the node is a 16000 ℎ power bank PB-T3 placed in a box with IP 67 protection.

3.3. Software

Main algorithm This software follows the classic client-server

model. The server is based on Raspberry Pi and Apache. It provides a set of customer services and shares its resources (files). The server waits to receive a user request, processes it, and returns the response. Clients are all devices that use the information provided by the user interface. The clients require specific services and control the incoming data. The link between the client and the server is the global network.

The main steps in the client - server communication are as follows:

• Web server and client connect via HTTP protocol; • The corresponding page is transmitted to the

browser as HTML code and is rendered; • The customer enters the data and sends the request

to the server; • Depending on the request, the server retrieves the

file and prepares a graph with the concentration for the selected period of time;

• The result of the executed query is received by the browser and is embedded in an HTML page. Its goal is to be interpreted by the client.

The device works under a free Linux operating

system – Raspbian, which is based on Debian and is optimized for the Raspberry Pi hardware. It is a highly customizable operating system and supports lots of different programing languages and environments. The

implemented software is written using Python and PHP high-level languages. Front-End development is done in HTML and CSS. The user interface is realized as a web page and is accessible from any Internet browser. This makes the system flexible and convenient for use.

The Back-End software could be provisionally divided into two parts. One is the software performing collection and processing of the data, and the other is the software processing the incoming requests to the server. For the server we have used Apache open source software, which supports multiple modules and extensions.

After a successful boot of the operating system, the Apache server is started and Python script executed. First, the script communicates with the peripheral modules. The time and date provided by the GPS RTC module are used to synchronize the system clock. This is necessary when there is no established Internet connection typically used for synchronization by Pi.

Fig. 3 shows a simplified algorithm for the program running on the server.

The program starts by initializing the OPC connected to the SPI interface. Depending on the system state and the current time, a new file is created or the entry continues in the previous file.

The data is written in the form of text files and a new file is created every 24 hours containing the , . and concentration values averaged per hour. The selected format is CSV and the data is in comma separated format. All files are available and can be downloaded through the web page.

The program continues to read the sensor values every second and store them in the temporary memory for a one-hour averaging.

On the basis of centration measurements made every second, charts are generated in the form of png images, which are updated every hour. This is done using the Python matplotlib module. Also, once an hour, the bar graphs with one-hour averages are updated. Further, the current day average pollution level is calculated in / .

Since the algorithm is an infinite continuous cycle, the generation of the graphics is moved to a separate thread in the program so that sampling the data is at regular intervals of one second without interruption.

GPS Data The station is mobile, and data about its location is

needed. In this way, when installing multiple nodes, you can clearly distinguish the data for different locations.

To realize this functionality, an integrated GPS module is integrated into the system.

“Е+Е”, vol. 53, 1-2, 2018 5

Fig. 3. Algorithm of the main program [14].

After loading the operating system, the information

received from the GPS receiver is read via the serial interface. The GPS module is equipped with a real-time clock with a reserved power supply. When a GPS signal is captured, the time and date are set and the station location data is refreshed, at a frequency of 1 . The GPS module packs the data in National Marine Electronics Association (NMEA). The NMEA GPS data is sent to the single-chip microcomputer where it is interpreted. Due to the lack of a built-in RTC, Raspberry Pi performs time synchronization over the internet with every start of the system. The station provides automatic time and date setting of the operating system by GPS data, allowing for proper operation even in the absence of Internet connectivity.

The altitude, latitude, longitude, and the Coordinated Universal Time (UTC) are displayed on the site as data for the corresponding sensor system. Geographical degrees are used for the coordinates and the altitude is expressed in meters.

In order to be able to report pollutant levels for a specific date, a functionality has been developed that enables the user to visualize the data on the web page for the , . and on both individual and combined graphs.

This is implemented using a PHP program running on the server. It creates an HTML dropdown menu from which the user can choose a date for which the graph should be drawn.

4. Experimental Results

4.1. The Complete mobile node

Fig. 4 shows the complete mobile node.

Fig. 4. View of the sensor node.

The node is placed in a plastic enclosure with IP 67 protection with GPS module and power bank. The components used in the sensor node and their approximate prices are given in Table 3.

Table 3 List of components of the sensor node

Part Commercial name Price, Euro

Main board Raspberry Pi 3 35 OPC-N2 400 GPS Module GPS NEO-7M 30 Enclosure - 15 Cables + batteries - 25

The cost of the components is about 505 Euro. A

comparison with the typical price of professional systems makes the mobile sensor node highly competitive.

4.2. Data acquisition

The data acquisition process is separated in different threads. One script reads the GPS data every second

6 “Е+Е”, vol. 53, 1-2, 2018

and obtains information about the location of the system and its altitude. Another Python script takes care of the OPC control. By means of commands sent through SPI, the microcomputer reads , . and

concentrations every second. Particulate matter data is processed and aggregated

for a period of 1 hour. A sample graph for instantaneous concentration is given in Fig. 5.

Fig. 5. Instant values measured for period of 1 hour.

The averaged information over every hour is stored

on a disk by a script running in another thread. Some examples of the diurnal variation in . and are shown in Fig. 6 and Fig. 7. All graphs for the past 24 hours and graphs for instant values for the past hour are available on the website. The data is stored in .csv (comma separated values) format as files per every day and is also available on the web.

Fig. 6. Diurnal variation in . mass concentration for Sofia, Kazichene.

Fig. 7. Diurnal variation in mass concentration for Sofia, Kazichene.

At the Back-End server there is a PHP code that

takes user requests from the web page and draws plots for selected past periods (Fig. 8). The raw .csv files can be downloaded and used for additional calculations.

Fig. 8. Twenty-four hour measurement of , . and values averaged for a period of 1 hour, Sofia,

Kazichene.

4.3. Validation of the measurements

In Europe, measurements of the concentrations of within national and international monitoring

networks have been carried out for more than 2 decades. Measurements of . have been performed over a shorter period of time and in limited locations because the European Air Quality directive on . target values has been introduced more recently. Monitoring in the period 1994–2008 showed that annual concentrations varied from 5 to about 50 / [15]. The . concentrations ranged from 3 to 35 / . In southern Europe the . -to-

ratio is between 0.4 and 0.6.

“Е+Е”, vol. 53, 1-2, 2018 7

As it can be seen in Fig. 8, the concentration of . and in the suburban area of Sofia (Kazichane) varies in the interval from about 3 to 20-22 / and the . to ratio is in the range between 0.3 and 0.7.

Due to the fact that the Air Quality Directive does not set any limit values for there are limited long-term measurements of particles of that size.

Field evaluation For the field evaluation the tested node was

collocated with the fixed station for air quality monitoring in Sofia Drujba. The concentration measured with the sensor node and the fixed authority station over a 24ℎ period is shown in Fig. 9.

The correlation coefficient between the sensor and the reference instrument is relatively high – 0.763. In particular, the sensor has detected two sharp increases in dust levels at 9 and 22 hours.

Fig. 9. Twenty-four hour measurement of values averaged for period of 1 hour near Drujba fixed station on

the 14th of April, 2017.

The long-term measurements with a duration of one week, which we have made, also show a good recurrence of the results and the correlational coeffi-cient between the sensor and the reference instrument is within the range of 0.72– 0.84.

5. Conclusion

Increasingly complex and demanding regulations to mitigate particulate matter pollution mean that aerosol measurements are becoming more and more time- and resource-intensive.

The continuous convergence of micro- and nano-technologies with the wireless communication net-works leads to significant paradigm shift in the field of

monitoring. The latest developments in the field of

sensor technologies have enabled the implementation of low-cost (in the 10-50 Euro segment) and semi-professional (in the higher price segment of 500-1000 Euro) sensors which are designed to be used in urban environments [17], [18].

The accuracy of low-cost sensors is still limited, and the devices implemented on their basis usually require significant calibration and maintenance costs [19], [20].

In this study, the design and implementation of a node for real time monitoring of basic parameters based on semi-professional sensor devices is presented. The results of the study show that they have significant advantages when compared to low-cost solutions. They can perform with acceptable precision in estimating

exposures over time and space. If properly calibrated, they could provide useful exposure estimates for health effects studies and support better assessments of the variation in relative to a central site monitor. The compact size of these monitors and the flexibility in data acquisition, communication and in- and outdoor localization can favor their wide application in field exposure studies.

The deployment of such systems in large numbers can provide near real time parameter information with higher space and time resolution. So the imple-mentation of this smart sensor node for on-line partic-ulate matter monitoring can enable the use of the ar-chived data in big data analysis and data mining.

Acknowledgments

We thank Assist. Prof. Kiril Filipov of the Labora-tory for Technical diagnostics and service of the Technical University of Sofia and Mr. Christo Angelov - of the Executive Environment Agency (ExEA) - Ministry of Environment and Water for supporting our field evaluation.

REFERENCES

[1] Directive 2002/49/EC of the European parliament and of the council of 25 june 2002 relating to the assessment and management of environmental noise. Off. J. Eur. Communities, vol. L189, p. 12–25, Jul. 2002.

[2] World Health Organization, "Review of Evidence on Health Aspects of Air Pollution. REVIHAAP Project: Technical Report. WHO,," Copenhagen, 2013.

[3] "Smart City-Strategie Berlin," Senatsverwaltung für Stadtentwicklung und Umwelt - Berlin, 2015.

[4] Yu Zheng, Furui Liu, Hsun-Ping Hsieh. U-Air: When Urban Air Quality Inference Meets Big Data. KDD’13, Chicago, USA, August 11–14, 2013.

[5] Viana, M. Ed. Urban Air Quality in Europe. Berlin Heidelberg, Springer-Verlag, 2013.

0

10

20

30

40

50

60

70

0 6 12 18 24

Conc

entr

atio

n, µg/m3

Time, h

Sensor nodeDrujba fixed station

8 “Е+Е”, vol. 53, 1-2, 2018

[6] Jovasevic-Stojanovic, M. et al. On the use of small and cheaper sensors and devices for indicative citizen-based monitoring of respirable particulate matter. Environmental Pollution, vol. 206, pp. 696-704, 2015.

[7] Budde, M., R. El Masri, T. Riedel, M. Beigl. Enabling low-cost particulate matter measurement for participatory sensing scenarios. 12th International Conference on Mobile and Ubiquitous Multimedia (MUM 2013), Lulea, 2013.

[8] European Environment Agency. Air quality in Europe report. Copenhagen, 2017.

[9] Impact Analysis. European Commission, 2013. [10] Williams, R., R. Long, M. Beaver et al. Sensor

evaluation report. Environmental Protection Agency, Office of Research and Development, Washington, DC, May 2014.

[11] Raspberry Pi Foundation. Raspberry Pi 3 Model B Documentation. [Online]. Available: https://www.raspberrypi.org/documentation/. [Accessed 2017].

[12] Wriedt, T., W. Hergert. The Mie theory: Basics and applications. Heidelberg, Springer, 2012.

[13] Alphasense Ltd. Alphasense user manual OPC-N2 optical particle counter. April 2015.

[14] Stoeva, T. System for environmental noise parameter measurement, Master’s thesis. Technical University of Sofia, 2017.

[15] Fuzzi, S. et al. Particulate matter, air quality and climate: lessons learned and future needs. Atmos. Chem. Phys., vol. 15, 2015, pp. 8217–8299.

[16] Sukwon Choi, Nakyoung Kim, Hojung Cha, and Rhan Ha. Micro sensor node for air pollutant monitoring: hardware and software issues. Sensors, vol. 9, 2009, pp. 7970-7987.

[17] Jiang, Y. K. Li, L. Tian et al. MAQS: A personalized mobile sensing system for indoor air quality monitoring. Proceedings of the 13th ACM International Conference on Ubiquitous Computing, Beijing, 17–21 September 2011.

[18] Bales, E., N. Nikzad, C. Ziftci et al. Personal pollution monitoring: mobile real-time air-quality in daily life. San Diego, 2014.

[19] Piedrahita, R., Y. Xiang, N. Masson. The next generation of low-cost personal air quality sensors for quantitative exposure monitoring. Atmospheric Measurement Techniques, vol. 7, 2014, pp. 3325–3336.

[20] European Environment Agency. AirBase - The European air quality database. 2016. [Online]. Available: http://www.eea.europa.eu/data-and-maps/data/airbase-the-european-air-quality-database-3.

[21] Hasenfratz, D., O. Saukh, S. Sturzenegger, L. Thiele. Participatory air pollution monitoring using smartphones. Proceedings of 2nd International Workshop on Mobile Sensing, Beijing, China, 16 April 2012.

[22] Hasenfratz, D., O. Saukh, C. Walser et al. Deriving high-resolution urban air pollution maps using mobile sensor nodes. Pervasive Mobile Comput., vol. 16, 2015, pp. 268–285.

Assoc. Prof. Dr. Marin B. Marinov received a M.Sc. de-gree in Technical Cybernetics from the Technical University of Ilmenau, Ilmenau, Germany in 1985 and a Ph.D. degree in Computer Science from the Technical University of Ilmenau, in 1989. Since 2001, he has been an Associate Professor at the Department of Electronics, Faculty of Electronic Engineering and Technologies, Technical University - Sofia, 8, Kliment Ohridski Blvd., 1756 Sofia, Bulgaria.

e-mail: [email protected].

Received on: 27.01.2018

THE FEDERATION OF THE SCIENTIFIC-ENGINEERING UNIONS IN BULGARIA /FNTS/

is a professional, scientific - educational, nongovernmental, nonpolitical, nonprofit association of legal entities - professional organizations registered under the Law on non-profit legal entities, and their members are engineers, economists and other specialists in the fields of science, technology, economy and agriculture.

FNTS has bilateral cooperation treaties with similar organizations in multiple countries.

FNTS brings together 19 national associations – Scientific and Technical Unions (STU), 34 territorial associations, which have more than 15 000 professionals across the country.

FNTS is a co-founder and member of the World Federation of Engineering Organizations (WFEO).

FNTS is a member of the European Federation of National Engineering Associations (FEANI), and a member of the Standing Conference of engineering organizations from Southeast Europe (COPISEE), UN Global Compact, European Young Engineers (EYE). The Federation has the exclusive right to give nominations for the European Engineer (EUR ING) title.

Contacts: 108 Rakovsky St., Sofia 1000, Bulgaria WEB: http: //www.fnts.bg E-mail: [email protected]

“Е+Е”, vol. 53, 1-2, 2018 9

Plasma based sterilization using new multipoint ignition accessory

Kiril Ivanov, Ivo Iliev, Serafim Tabakov

This article discusses the physical phenomenon of plasma and its basic principles. The processes

of the plasma generation as well as the various particles that are produced with it are described. An overview of the effects of each generated during plasma particle on the processes of sterilization and disinfection is made. A solution for the mechanical part of plasma sterilizer accessory is presented. It is suggested a possible operating principle of such a sterilizer, how to control the mechanical part of the shown sterilizer. The main topic of the article is based on a series of experiments. The results of these experiments will evaluate the effect of each RF signal parameter on the plasma generated by the same signal.

Keywords – Accessory, Disinfection, Plasma, Sterilization.

Плазмена стерилизация, чрез нов многоточков аксесоар (Кирил Иванов, Иво Илиев, Серафим Табаков). В настоящата статия е разгледано физичното явление плазма и неговите основни принципи. Описани са процеса на генериране на плазмата и различните частици, които се получават при него. Направен е кратък обзор на въздействието на плазмата, върху процесите на стерилизация и дезинфекция. Показано е примерно решение за механичен накрайник за плазмен стерилизатор. Предложен е възможен принцип на работа и управление на механичната част от показания стерилизатор. Представени са редица експерименти, чиято цел е да се оцени въздействието на различните параметри на изходен високо-честотен сигнал върху плазмата генерирана с негова помощ.

Disinfection and sterilization are both decontami-nation processes. While disinfection is the process of eliminating or reducing harmful microorganisms from inanimate objects and surfaces, sterilization is the pro-cess of killing all microorganisms. Sterilization also destroys the spores of various organisms present on surfaces, in liquids, in medication or in compounds such as biological culture media. Such "extreme" forms of decontamination are needed during surgery or in environments like industrial, laboratory or hospi-tal. In everyday life it is more practical to use disinfec-tion [12].

One newly used methods for sterilization and dis-infection is based on the use of plasma and plasma products. Plasma is often called the "Fourth State of Matter". Although found in virtually every home and business, gas plasma is not well known. The term plasma was first introduced by chemist Irving Lang-muir in the 1920s. He proposed the following descrip-tion: “A plasma is a quasi-neutral gas consisting of positively and negatively charged particles (usually ions and electrons) which are subject to electric, mag-netic and other forces, and which exhibit collective behavior”.

Plasma can be simply considered as a gas of charged particles. These charged particles are nega-tively charged electrons and highly charged positive ions, being created by heating a gas or by subjecting gas to a strong electromagnetic field. However, true plasma production is from the distinct separation of these ions and electrons that produces an electric field, which in turn, produces electric currents and magnetic fields [2].

A significant number of highly charged particles together make plasma electrically conductive so that it responds strongly to electromagnetic fields [3].

Depending of the temperature, there are two types of plasma – high and low temperature plasma. High temperature plasma is more often on the Earth, most often as a lightning [10], [11].

Low temperature plasma is used for surface etch-ing, disinfection, sterilization etc. It is mostly created by ionization of gases in vacuum or low pressure envi-ronment, inside chambers with controlled environ-ment in them – atmospheric gases are absent or have very low concentration.

10 “Е+Е”, vol. 53, 1-2, 2018

Brief overview of recent research plasma-based sterilization/decontamination methods is given [1], [4].

Plasma-based sterilization/decontamination.

Biological materials can be exposed to plasma in two different methods: (1) “direct exposure” and (2) “remote exposure”. “Direct exposure” is when the sample to be treated is in direct contact with the plas-ma [13]. All plasma-generated agents, including charged particles, are in contact with the sample. In the “remote exposure” method, the sample is placed at a distance from the plasma volume or in an adjacent chamber. In this configuration, the amount of heat transmitted to the sample is reduced. Charged parti-cles do not play a role since they recombine before reaching the sample, and many of the short-lived neu-tral reactive species also do not reach the sample. In the following section contribution of the four main inactivation factors of a non-equilibrium high-pressure air plasma are reviewed.

Heat: Heat-based sterilization methods use either moist heat or dry heat. In the case of moist heat, such as autoclave, a high temperature and high pressure is used. Dry heat sterilization requires temperatures close to 170˜ and treatment times of about 1h.

UV Radiation: UV radiation in the 200-300nm wavelength range with doses of several mW·s/cm2 causes lethal damage to cells. This effect is strongest for UV light at 285nm wavelength. Amongst UV ef-fects on cells of bacteria is the dimerization of thy-mine bases in their DNA strands. This inhibits the ability of the bacteria to replicate properly [6], [8].

Charged particles: Charged particles play a very significant role in the rupture of the outer membrane of bacterial cells. Electrostatic force caused by charge accumulation of the outer surface of the cell mem-brane could overcome the tensile strength of the membrane and cause its rupture.

When charged, a body of the size of a bacterial cell (in the µm range) experiences an outward electrostatic force because each charge is subjected to the repulsive forces of all the similar changes accumulated on the cell surface. This force is proportional to the square of the charging potential, Φ, and inversely proportional to the square of the radius of the curvature of the sur-face, r. Therefore, the smaller the radius of curvature the stronger the electrostatic force. The charging po-tential Φ depends on the ratio of the ion mass to the electron mass. So gases with larger atomic mass lead to higher electrostatic forces [7]. Based on this the condition for membrane disruption is:

(1) | | > 0.2 ∙ ∙ ∆ / ∙ /

where r is the radius of the curvature, Δ is the thick-ness of the membrane, and Ft its tensile strength [5], [9].

Reactive Species: In high-pressure non-equilibrium plasma discharges, reactive species are generated through various collisional pathways, such as electron impact excitation and dissociation. Reactive species play an important role in all plasma-surface interac-tions. Air plasmas, for example, are excellent sources of reactive oxygen-based and nitrogen-based species, such as O, O2, O3, OH, NO, NO2, etc.

Disinfection/Sterilization accessory proposal

Based on the information above a new disinfec-tion/sterilization accessory could be designed. For the accessory is important to define whether the plasma is in direct contact to the treated surface or not (direct exposure or remote exposure method). Since the steri-lization/disinfection effect of the heat is proven, the appropriate method should be with direct exposure. Considering the relation between plasma spot, energy consumption and generated temperature, we propose a conception for accessory (Fig. 1) including many small plasmas that will be ignited in series to form one big plasma spot.

Fig. 1. Example structure for Plasma based sterilization

accessory.

Structure like this should have as much as possible active electrodes (pins). Each electrode will create plasma to the treated surface. This surface could be referenced to earth or isolated. Electrodes should be activated in some series to create “moving” plasma along the accessory. This ignition sequence could be electrically or mechanically controlled. The so called

“Е+Е”, vol. 53, 1-2, 2018 11

spacer guarantees the stable (not changing) distance between active pins and surface. If the distance is always same, the energy delivered by the plasma for 1s (or for some other predefined period) will be con-stant. This means that the heat, the UV light and the charged particles generated for the respective period will be constant. This way the device will guarantee the same behavior during every procedure. To deter-mine the length of the spacer the ignition distance in various conditions should be measured.

Ignition distance measurements

Ignition distance is the distance between the active and return electrodes, when the plasma is stable. In addition, one more parameter should be taken into consideration – strimmer length. This is the minimum distance between the active electrode and the treated surface when the ionization of the medium gas medi-um for the plasma is visible.

To perform the tests described below in the article, a special device is used. It is based on existing RF technology used in electrosurgical equipment. Im-portant feature of this device is the capability to deliv-er controlled gas flow to the working area. Any type of gas, stored in a tank under pressure, could be used. The combination of RF output signal and delivered gas is the foundation for plasma generation. Test setup used for the next experiments is shown on Fig. 2.

Fig. 2. Test set-up used for performing following

experiments.

The device (RF Unit) has built in sensors for measuring real time RMS (Root Mean Square) Volt-age and RMS Current. Based on the measurements of these sensors and on the internal schematic of the device, following parameters are controlled:

• Maximum RMS Voltage – output voltage can be controlled from 0V to 1200V;

• Maximum RMS Current – output current can be controlled from 0A to 1.6A;

• Output power – calculated based on the voltage and current measurements. Module could deliver

up to 100W output power; • Frequency – output signal frequency can be se-

lected in the range from 320kHz to 500kHz; • Gas flow – gas delivered to the working area can

be adjusted in the range from 1slpm to 5slpm (“slpm”- Standard Liter Per Minute). Here should be noted that the gas used in the following experiments is helium.

Special software developed for this module con-trols all output signal parameters during the tests.

Maximum output RMS Current: Purpose of the first test was to evaluate the effect of the maximum output RMS current on the ignition distance and strimmer length. In this test all other parameters were constant:

• Maximum RMS Voltage – 600V; • Frequency – 336.13kHz; • Duty cycle – 4 ON pulses, 12 OFF pulses; • Gas flow – 4slpm. Results from the experiment are shown on Fig. 3.

Fig. 3. Ignition distance and strimmer length depending on

the maximum output RMS current.

In the focus of next experiment was the evaluation of interdependency between RMS current and plasma impedance when the ignition distance and the strim-mer length are constant. Results are shown on Fig. 4.

Fig. 4. Plasma impedance depending on the maximum

output RMS current.

12 “Е+Е”, vol. 53, 1-2, 2018

Maximum output RMS Voltage: To evaluate the effect of the maximum output RMS voltage on the ignition distance and strimmer length the next test was performed. In this test, other parameters are:

• Maximum RMS Current – 1.25A; • Frequency – 336.13kHz; • Duty cycle – 4 ON pulses, 12 OFF pulses; • Gas flow – 5slpm. Results are shown on Fig. 5.


the maximum output RMS voltage.

Gas flow rate: Subject of the next test was the evaluation of ignition distance and strimmer length depending on the gas flow rate. In this test, other parameters are:

• Maximum RMS Current – 1.25A; • Maximum RMS Voltage – 600V; • Frequency – 336.13kHz; • Duty cycle – 4 ON pulses, 12 OFF pulses. Results are shown on Fig. 6.


the gas flow rate.

Duty cycle: Subject of the next test was the evaluation of ignition distance and strimmer length depending on the duty cycle. In this test, other parameters are:

• Maximum RMS Current – 1.25A; • Maximum RMS Voltage – 600V; • Frequency – 336.13kHz; • Gas Flow – 4slpm. Results are shown on Fig. 7.


the duty cycle.

Frequency: For the next test the focus is on the evaluation of ignition distance and strimmer length depending on the output signal frequency. In this test, other parameters are:

• Maximum RMS Current – 1.25A; • Maximum RMS Voltage – 600V; • Duty cycle – 4 ON pulses, 12 OFF pulses; • Gas Flow – 4slpm. Results are shown on Fig. 8.


the output signal frequency.

To better understand the relation between the plasma parameters and the output signal frequency additional experiment is performed. In this experiment measurements on the plasma impedance depending on the output signal frequency are performed. Results are shown on Fig. 9.

“Е+Е”, vol. 53, 1-2, 2018 13

Fig. 9. Impedance of the plasma depending on the output

signal frequency.

For further evaluation of the influence of the output signal frequency on the characteristics of the unit, output signal waveform is observed. This experiment is performed with disabled feedback control of the unit and over 2000Ω non-inductive load. This value is selected based on the measured impedance of the plasma at 10mm distance. This is the typical distance between the active electrode and the treated surface allowing to have enough visibility. First experiment is performed at set for approximately 20W output power. Second one is performed at set for approximately 40W.

Results from this test provides information about the correlation between the CF of the output signal and output power set and output signal frequency. Other parameters are:

• Duty cycle – 4 ON pulses, 12 OFF pulses; • Gas Flow – 4slpm. Crest factor is calculated using the formula:

, where: IPEAK is the maximum value of the current

and IRMS is the RMS value of the current. Results are shown on Fig. 10.

Fig. 10. Calculated CF (crest factor) of the output signal

depending on the frequency.

Above results proves that the CF depends on the frequency and doesn’t depend on the output power.

To evaluate only the influence of the frequency (not the crest factor) on the ignition distance and the strimmer length, additional two tests are conducted. The RF device is set to deliver output signal with peak-to-peak voltage approx. VPP=5.00kV and for the second test with RMS voltage approx. VRMS=700V. Other set parameters are:

• Duty cycle – 4 ON pulses, 12 OFF pulses; • Gas Flow – 4slpm; • OPEN LOOP – ON; • SMPS Range – ON; • VPP – 5.00kV; • VRMS – 700V. Results are shown on Fig. 11 and Fig. 12.

Fig. 11. Ignition distance and strimmer length depending

on the frequency at VPP 5kV.

Fig. 12. Ignition distance and strimmer length depending

on the frequency at VRMS 700V.

Results interpretation

Based on all of the data presented in previous section (Ignition distance and Strimmer length) some conclusions are obtained:

• Increasing the maximum RMS current does not affect the ignition distance and strimmer length.

14 “Е+Е”, vol. 53, 1-2, 2018

This was expected result because these two parameters are mostly affected by the ionization of the medium gas, which depends mostly on the peak-to-peak voltage of the output RF signal. However increasing the RMS current flowing thru the plasma reduces the impedance of the plasma beam. The higher the RMS current is the more charged particles are induced in the plasma beam.

• Increasing the maximum RMS voltage of the output RF signal increases both the ignition distance and the strimmer length of the plasma. This is caused by the better ionization of the medium gas, because of the increased peak-to-peak voltage of the output signal.

• Increasing the duty cycle of the output signal is leading to decrease of both the ignition distance and strimmer length. The increase of the duty cycle (while the RMS of the voltage is the same) leads to decrease of the peak-to-peak. As already proven a decrease in the peak-to-peak value is decreasing in both ignition distance and strimmer length.

• Gas Flow is the first parameter of the plasma that shows some irregularity in the test results. As visible from the results the increase of the gas flow is leading to increase of the ignition distance and strimmer length until some a specific point. After that, the quantity of the medium gas molecules become too high to be ionized and the flow is able to cool down the ionized gas, leading to reduced ignition distance and strimmer length.

All results from the tests performed by changing the output signal frequency are not consistent. It could be conclude that the frequency of the output signal has very strong influence over the parameters of the plasma. In addition, the frequency is affecting the parameters of the final output stage of the RF signal generator. These effects should be subject to other experiments.

Acknowledgement

The present work was supported by project DN 07/21 FUND “Scientific Investigations”, Bulgaria.

REFERENCES

[1] Laroussi, M. Low temperature plasma-based sterilization. Wiley-YCH Verlag, 2005.

[2] Gallagher, P.T. Introduction of plasma physics. Trinity College Dublin, 2003.

[3] Schwartz, S. Basic plasma physics, Imperial College London, 1993.

[4] Lerouge, S., M.R. Werheimer. Plasma sterilization: a review of parameters, mechanisms, and limitation. 1999.

[5] Mendis, D.A., M. Rosenberg, New J Phys. 2003, 5, 41.1.

[6] Norman, A. J. Cell. Comp. Physiol. 1954, 44, 1. [7] Mendis, D.A., M. Rosenberg, F. Azam. IEEE

Trans. Plasma Sci. 2000, 28, 1304. [8] Laroussi, M. IEEE Trans. Plasma Sci. 2002, 30,

1409. [9] Stalder, K.R. ‘Plasma characteristics of

electrosurgical discharges. Proc. Gaseous Electron. Conf., AIP, San Francisco, CA 2003, p. 16.

[10] Herrmann, H.W., I. Henins, J. Park, G.S. Selwyn. Phys. Plasmas 1999, 6, 2284.

[11] Yamamoto, M., M. Nishioka, M. Sadakata. Proc. 15th Int. Symp. Plasma Chem., Vol. II, University of Orleans, Orleans 2001, p. 743.

[12] Low temperature plasma sterilizer, https:// tuttnauer.com/medical-autoclaves/low-temperature-sterilizer

[13] PJ - bench top plasma unit, http://www. astp.com/plasma-pj

Eng. Kiril Sl. Ivanov graduated master engineer

tel.:+359 887000654 е-mail: [email protected]

Prof. Ivo Ts. Iliev, DSc is Head of the Lab. of Biomedical Engineering, Department of Electronics at the TU-Sofia. Areas of scientific research – Medical instrumentation and biomedical signal processing.

e-mail: [email protected]

Assoc. Prof. Serafim D. Tabakov is a lecturer and a leading researcher in the Lab. of Biomedical Engineering, Department of Electronics at the TU-Sofia.



“Е+Е”, vol. 53, 1-2, 2018 15

Design of axial rotary wireless energy transmitter

Valeri Petkov, Nikolay Madzharov

This article provides a review of the possible types of rotary wireless transmitter (RWT)

constructions. It also presents a design methodology of transmitting and receiving coils for axial RWT, providing maximum energy efficiency with minimal dissipation losses. To reduce the impact of air gap, between the transmitter and the receiver, the coils are connected in resonant circuits. The design of resonant circuit elements and their matching with the high frequency generator are also investigated. A mathematical analysis of the transmitter and the receiver coils is presented as separate inductive elements and as interconnected ones by the electromagnetic field. There are four main types of RWT construction introduced in this article, examining in details just one of them – the axial one. In order to clarify the advantages and the disadvantages of the construction, computer simulation was made, showing the distribution of the magnetic field and its intensity at the individual points of the wireless transmitter.

Keywords – Electromagnetic analysis, Matching, Resonant circuit, Rotary transmitter, Winding configurations, Wireless transfer

Проектиране на аксиален ротационен безконтактен предавател на енергия (Валери Петков, Николай Маджаров). Тази статия представя преглед на възможните видове конструкции на въртящи се безконтактни предаватели на електрическа енергия. Също така предлага методика за проектиране на предавателна и приемна намотки за аксиален ротационен предавател, осигурявайки максимална енергийна ефективност с минимални загуби на разсейване на полето. За да се намали влиянието на въздушната междина между предавателя и приемника, намотките се съгласуват чрез резонансни кръгове. Изчисляването на елементите от резонансните вериги и тяхното съгласуване с високочестотния генератор също е обект на представената методика за проектиране. Направен е електромагнитен анализ на предавателната и приемната намотки, като самостоятелни индуктивни елементи и като взаимосвързани чрез електромагнитното поле. Статията представя четири основни вида конструкции на ротационни безконтактни предаватели, като разглежда основно структурата и параметрите на една от тях – аксиалната. За да се изяснят предимствата и недостатъците на предложената конструкцията, е извършена компютърна симулация, показваща разпределението на магнитното поле и неговата интензивност в отделните зони на безконтактния предавател.

Introduction

Wireless power transmission (WPT) is an innovative technology which launches a new stage in the development of automated systems. It also involves the optimization of a lot of technological processes, as it offers great flexibility, because of the removal of power cables, the lack of moving electrical contacts and the associated problems (sparks, wear, etc.). It also allows application in explosive environments, such as high humidity and high speeds. It is used for the wireless supply of consumers, which perform linear and rotary motion.

The rotary motion is used in almost all areas of the industry. It is often necessary to transfer energy to the

rotating elements and this requires use of a rotary transmitters. It can be realized by sliding contact rings or a RWT. The article discusses the possible constructions of RWT, as the emphasis was put on the construction of axial transmitter. It is necessary to achieve the optimization of the RWT and improve its efficiency. This is related to the reduction of losses from the self-inductance, the improvement of the magnetic coupling ratio and the creation of conditions for easy matching with the other functional blocks of the WPT system. The mutual magnetic induction M and the self-inductivities L1, L2 are of great significance to improve the efficiency. Their values depend upon the geometry of the coils - the number of windings, the

16 “Е+Е”, vol. 53, 1-2, 2018

distance between them, the presence of magnetic concentrators, the winding style and the size of the air gap.

The article presents an electromagnetic analysis of the transmitting and receiving coils and on this base is introduced a mathematical model of designing a rotary wireless transformer. A computer simulation was made, showing the advantages and the disadvantages of the axial RWT design.

Types of rotary wireless transmitters

There are two basic RWT constructions - axial (A) and radial (R) shown on Fig.1a and Fig.1b. They are divided into several sub types, differing from the location of the coils and the shape of the body to which they are attached. A variation of the two constructions is the coaxial transmitter - Fig.1c and Fig.1d. It combines the advantages of both configurations.

a)radial ; b)axial ; c) coaxial R ; d) coaxial A

Fig.1. Types of rotary wireless transmitters.

In the design of the conductors for the individual types of RWT, a major problem is the inconsistency between the standard ferrite cores, offered by the companies and the dimensions of the transmitter, designed by the constructors. This requires the usage of composite magnetic cores, which complicates the design and construction of the mechanical system. A solution the problem is use of new magneto-dielectric materials, which have a number of advantages, comparing to the traditional ferrites [1]. Table 1 presents a comparison between the most commonly used materials for making transformer magnetic cores.

Table 1 Types of materials

Material Perme- ability

El. Re- sistivity Ohm.cm

El.Strength Vrms/mm (300kHz)

Availab. in sizes and shapes

Curie temp. 0C

Ther. cond. W/cm2

Fer.559 18-20 High >15000

Medium <100

Any when machined >300 0.04

Fer.119 7 Very High

>300 Vrms/mm

Any when machined >300 0.02

Fluxtrol 50 50 Low <100

Vrms/mm Any when machined >300 0.05

Fluxtrol HF 5 Very

High >> 300 Vrms/mm

Any when machined >300 TBD

The conclusion that could be made is the magneto-dielectric materials have a significant advantages, because they are subject to mechanical machining and practically each and one required shape of a magnetic core could be obtained.

Rotary transmitter with axial (A) configuration

Cross-section of an A RWT is shown in Fig.2 [2]. The configuration is applicable in cases with constructive limitation of the permissible dimension in direction x, according to Fig.2. This requires an extension of the transmitter in direction y, i.e. parallel to the axis of rotation, which at the same time improves the magnetic coupling factor. It is typical for this type of transmitters to allow instability, fluctuation and slight displacement in the y - direction, making this kind of rotary transmitter suitable for hit- drilling purposes. A disadvantage of the axial RWT is the construction difficulty, which appears in producing the transmitter coil.

Fig.2. Axial rotary wireless transmitter.

Electromagnetic analysis

The main design symbols and dimensions are as follows:

Lw,Wr,Wt – geometric dimensions of the coils, length and width, as the index denotes transmitter (t) and receiver (r);

D1 – the internal diameter of the receiver. It is determined by the axis of rotation value. Often this is a spindle axis or an axis of a certain type electric motor.

D2 – diameter defines the section of the receiver; D3 – the external diameter of the receiver; D4 – the internal diameter of the transmitter; D5 – diameter defines the width of the transm. coil; D6 – the external diameter of the transmitter; C – length of RWT; Diameters D5 and D6 define the cross section of the

transmitter. The size of the air gap is determined by the diameters D3 and D4. The optimization of the Lw, Wr(t)

“Е+Е”, vol. 53, 1-2, 2018 17

and C allows improving the magnetic coupling factor between the transmitting and receiving coils, and reduce their parasitic parameters.

a, b, c, d, e, f, – geometric dimensions of the respective parts of the core; δ - air gap;

R – magnetic resistance of the individual parts, which set up the RWT chassis. They are a closed circuit with lengths l(n), as the intensity H of the field through them is constant. The sum of the individual lengths gives the average length of the magnet core.

(1) = ∑ = + + + + + + 2

The value of R in the individual parts is defined by:

(2) = , where

l(n) – length of a respective part; S(n) – section of a respective part; μM – magnetic permeability of the used material. The magnetic resistivity of the magnetic core RM is

the sum of the resistance of the individual parts:

(3) ( )= nM RR .

The influence of the air gap δ is determined by its magnetic resistance Rδ, according to the expression:

(4) S

lR

0μ= δ

δ ,

where lδ – length of the air gap; S – cross section of the air gap area.

The full magnetic impedance for the entire circuit of the RWT is presented in Fig.3 [3].

Fig.3. Equivalent scheme of the magnetic circuit.

The equivalent magnetic impedance of the closed magnetic circuit is defined by the sum of the magnetic resistances of the sequentially joined parts.

(5) δ+= RRR M

The current flowing through the transmiter coil creates a magnet flux Φ. The main part Φ0 passes through the magnet core and the other part generates a dissipation flux Φσ, i.e.

(6) σΦ+Φ=Φ 0 .

The absolute magnetic permeability of the material of the ferrite core exceeds in many times the magnetic permeability of the air ( μ >> μ0), therefore Φ0 >> Φσ and Φ ≈ Φ0.

The value of the MMF (Magneto-Motive Force) Fm is obtained by the following expressions:

(7) = . ,

(8) = Ψ . ,

where wt – windings of the transmitting coil; it – magnetic current through the transmitting coil.

This magnetic circuit could be presented by the following expressions:

(9) = Ψ . = Ψ . + = Ψ . +

For designing the coils, except the coil inductance L, another important parameter, which must be observed is the parasitic capacitor Cp - called coil capacitor [3]. It occurs between adjacent turns and between the rows of the coil. This capacitor induces additional losses in the coils, therefore it should be minimal. Segmentation of the coil is used in order to reduce Cp. From a construction point of view, decreasing the width of the coil Wr, Wt, respectively its rows, decreases Cp. Increasing Lr, Lt allows achieving the desired inductance at fewer rows of the coil. In addition, this parameters improves the magnetic coupling factor.

The voltage of the two coils of the RWT can be expressed by applying the Kirchof’s law, through their self inductance of the coils and their mutual inductance:

(10) = + = + , .

Electrical parameters of the transmitter and receiver coils

The coils of the WPT systems are inductive elements with self-inductance L. The equivalent scheme is shown in Fig.4 [3].

a) b)

Fig.4. An equivalent circuit of the inductive element: a) complete circuit, b) summary circuit.

The elements in Fig.4 are: L – self-inductance of the coil; Cp – a parasitic capacitor, which is determined by the self-capacitor CL, the capacitor between the

18 “Е+Е”, vol. 53, 1-2, 2018

windings of the coil CPW, the capacitor between the windings and the base CPB, and the capacitor of the connecting wires CPC.

The value of Cp can be determined by the generalized expression[3]:

(11) = = + = +

ξPW – dielectric permeability of the insulating material between the windings; ξPB – dielectric permeability of the coil base; ξ0 – dielectric permeability of vacuum; SC – cross section of the conductor; d – distance between the individual winding; lg – length of the air gap. RC – active resistance that reflects losses in parasitic

capacitor. Its impact grows, when the frequency gets higher, as a result the value of the quality factor decreases.

(12) = . [Ω] RFe – active resistance that reflects losses in the

magnet core; RCu – active resistance that reflects losses in the conductor.

The summary equivalent circuit of inductive element, which is used for the analysis of electromagnetic processes is presented in Fig.4b. It is a bipolar element with impedance:

(13) = . //

After simplification the expression is obtained:

(14) = . . /. . .

The inductive element could be considered as a resonant circuit with parallel resonance [4, 5]. The impedance acquires an active character in resonance, i.e. the imaginary part is equal to 0.

(15) 0.

.12

2 =−ω−L

CRCpL pL

The resonant frequency of inductive element fSRF (Self Resonance Frequency) is determined by (15).

(16) = = . − = .

The resonance frequency fSRF is an important parameter. To achieve optimality, the inverter operating frequency ω, must be well below than the resonant fSRF. For this purpose, the parameter relative

detune is introduced:

(17) =

Applying equation (16), the relative detune is obtained:

(18) = . .

The quality factor QL reflects the active loss in an inductive element. It represents the relation of the imaginary part of the impedance to the real part.

(19) = = .

The quality factor is frequency dependent and decreases, when the operating frequency is close to the resonant (fSRF) of the inductive element.

The inductive element is temperature dependent and its stability is quantitatively determined by the temperature coefficient TCL (Temperature Coefficient of inductance).

(20) = = ∆.∆ ° . 10 , [ /] The temperature stability is depends on the

construction of the inductive element and the materials used. In case of a magnetic core is significant the temperature stability of the magnetic permeability of the core - αμ.

Types of winding configurations

Different types of configurations are used to construct the coils, as each characterized by specific properties and parameters [6]. Possible ways of winding are demonstrated on the receiver part – Fig.5.

a) b)

c) d)

Fig.5. Configurations of coils: a) one-line coil; b) multi-line coil; c) multi-line spiral coil; d) biphilic coil.

One-line coil – has the lowest parasitic capacitor and hence, the highest resonant frequency. The disadvantage of this configuration is that it is

“Е+Е”, vol. 53, 1-2, 2018 19

impossible to reach a large inductance and in most cases it is necessary to increase the size of RWT or reduce the cross section of the conductor.

Multi-line coil – has the highest parasitic capacitor and the worst frequency properties. Allows to reach large inductances.

Multi-line spiral coil – has better frequency properties than the multi-line coil.

Bifilar coil – windings of a certain layer are placed in the next layer. This reduces the parasitic capacitor and allows to greater inductance. The rows number should to be even.

Design of axial RWT

The number of windings of the primary and secondary coils are determined according to the expressions:

(21) BA

tUw E

ΔΔ=..

min1 , 12 .w

U

Uw

E

out= ,

where Аmin – minimum cross-section of a magnetic core; B – magnetic induction; ∆t – RWT operating time, which is set by the operation mode of the converter – one-way or two-way mode [7]; UE – power supply voltage; Uout – output voltage.

Cross-section of the axial transmitter is shown on Fig.6. The following expression is used to determine the minimum cross-section of the magnetic core:

Fig.6. Cross-section of the axial transmitter.

(22) )( 21

2212min rrSSA −π=−= .

The inductance of a magnetic circuit is related to magnetic resistances R and the number of windings w. They could be measured with a LCR-meter or defined by the expressions:

(23) R

wL 1

1 = , R

wL 2

2 = , R

wwM 21.= .

The magnetic coupling factor k is related to the mutual inductance and self coil inductances:

(24) 21.LL

Mk =

To determine the DC resistance of the coil, are used the expressions:

(25) Sw

lcRDC .

.ρ= ,

where ρ - the specific resistance of a conductor material; S - conductor cross section; lC - the length of the conductor.

The length lC could be defined by the following expression:

(26) = ∑ 2. . ,

where n – number of layers; ri –radius of the corresponding row of the coil.

The skin effect increases the resistance of the coil in high frequency. It could be define by the expression bellow:

(27) )1.(

. / δ−−δ=

SDCSe

SRR ,

(28) f..μπ

ρ=δ , 0.μμ=μ r ,

δ – penetration depth of the skin effect; μ0 – magnetic permeability of the environment; μr – magnetic permeability of the conductor. The quality factor of the coil, is determined by the

expression:

(29) SRLQ /ω= .

In the design process is selected the topology of compensation. The compensating capacitors and the inductance of the coils make a resonant circuit. Serial-serial (SS) topology is used in the presented work [5]. The highest efficiency is achieved when the transmitting and receiving circuits are adjusted in resonance [8], [9], i.e.

(30) = = . = . .

The equation (30) is used to determine the compensating capacitor C1:

(31) 1201 ./1 LC ω= , = 2. . f..20 πω = – resonant frequency; f –

inverter control frequency. To define the compensating capacitor C2 is used the

expression:

(32) 2202 ./1 LC ω= .

20 “Е+Е”, vol. 53, 1-2, 2018

The current density through the coils and cross section of the conductor are determined by the following expressions:

(33) 21

21 .I

w

wI dens = ,

where I2 – maximum current value in the receiver. The effective current value for each of the coils are:

(34) ][,2

11 A

II dens

eff = ; ][,22

2 AI

Ieff = .

The cross section of the conductor (litz wire) is determined by the following expression.

(35) ][,.13,1 mmI

d effii Δ

= ,

where ∆= 2 ÷ 4, [ / ] – selected current density. The choice is made according to the design and the conditions of cooling of the RWT structure.

Experimental results

Тhe pictures in Fig.7 shows the designed axial transmitter. The core is made of ferrite E25 (Epcos). The pieces are mechanically processed afterwards and arranged fan-shaped. The ferrites are coated with epoxy resin. This provides mechanical strength and thermal performance.

The system for wireless power transfer is constructed with a half-bridge resonant inverter powered by 24V power supply. The Fig.8 shows a picture of the experimental mock-up.

The efficiency of the designed axial RWT has been investigated. The experiment was performed at a fixed load R=4,7Ω and the results are presented in Table 2.

Table 2 Experimental results

Uin Iin Pin Uout Iout Pout η[%] 24,1V 1,05A 25,3W 10,2V 2,1A 21,3W 84

a) b)

Fig.7. Axial wireless transmitter: a) transmitting part; b) receiving part.

The efficiency of the designed axial transformer is 84%. It could be optimized, if the air gap δ is reduced. In this case, a rotating mechanism must be balanced, without instability and displacement.

Fig.8. Axial RWT set up.

.

“Е+Е”, vol. 53, 1-2, 2018 21

The results of a computer simulation of axial RWT are represented in Fig.9. The analysis of the results, shows that the intensity of the magnetic field among the receiving coils is greater, which creates conditions for saturation of the magnetic core and reduce the efficiency.

Fig.9. Computer simulation of axial RWT.

Conclusion

The article examines the basic parameters of axial RWT. It introduces the methodology of electrical design and gives a detailed description of the electromagnetic processes by simulation. The main constructive requirements and limitations for the design of transmitting and receiving coils are given. Using the presented analysis, the axial transformer has been realized, and it was connected to a WPT system. The efficiency of the implemented system has been investigated experimentally. The presented information allows design and optimization of such systems with application in automation and in the high technology industry.

REFERENCES

[1] http://fluxtrol.com/ [2] Tosi, M., N. Bianchi, H. Herzog. Rotary

transformer design for brushless electrically excited synchronous machines. University of Padova, Technische Universitat Munchen, 2014, pp. 07-11, pp. 25-32.

[3] Jow, U., M. Ghovanloo. Design and optimization of printed spiral coils for efficient transcutaneous inductive power transmission. IEEE Transaction on Biomedical Circuits and Systems, vol.1, 3, September.

[4] Vuchev, Al., N. Bankov, G. Terziyski. Analysis of LCC resonant DC-DC converter with capacity filter in the load circuit. XLV International Scientific Conference on Information Communication and Energy Systems and Technologies ICEST 2010, Bitola, Macedonia, Proceedings of Papers, Vol. 2, 23-26 June 2010, pp.845-848.

[5] Lomonova, А., L. Encica, C. Smeets. Comparison of winding topologies in a pot core rotating. Eindhoven University of Technology, 2010 IEEE, pp. 103-110.

[6] Valchev, V., A. Bossche. Inductors and transfor-mers for power electronics. CRC press, 2005/3/24.

[7] Kraev, G., N. Hinov, D. Arnaudov. Multiphase DC-DC converter with improved characteristics for charging supercapacitors and capacitors with large capacitance. Annual Journal of Electronics, V6, B1, TU- Sofia, Faculty of EET, 2012, pp.128-131.

[8] Boeij, J.D., E. Lomonova, J. Duarte and A. Vandenput. Contactless energy transfer to a moving actuator. Conference Record of the 2006 IEEE Industry Applications Conference Forty-First IAS Annual Meeting, Tampa, FL, 2006, pp. 2020-2025.

[9] Nedyalkov, I., D. Arnaudov, N. Hinov. Studying of converter for charging and management of energy flows in photovoltaic system. Annual Journal of Electronics, 2015, pp.216-219.

PhD mag. Eng. Valeri P. Petkov - graduated industrial electronics in Technical and Electrical High School “General Ivan Bachvarov” Sevlievo, Bulgaria in 2009. He received his MTech degree in Electronics from Technical University of Gabrovo, Bulgaria in 2015. He has been involved in various research activities in the area of power electronics, induction heating, power supply systems and ultrasonic power converters. He is PhD student in Technical University of Gabrovo since 2015. His research areas are industrial electronics and systems for wireless power transfer.

tel.:+359 883 488 691 е-mail: [email protected]

Prof. Nikolay D. Madzharov - Dr. Thesis: Investigation and development of the autonomous inverters with energy dosing for electrothermics - Technical University of Gabrovo, Bulgaria, 1999. Research areas: Power and industrial electronics, Contact less systems 2004-2017. Lecturer at the Technical University of Gabovo since 2004 – until now. Article: more than 90; Books: 8; Patents: 3.

tel.:+359 887 991 701 е-mail: [email protected]


22 “Е+Е”, vol. 53, 1-2, 2018

Current waveform analysis for arc control in

electrosurgical devices

Viktor Tomov, Ivo Iliev, Vessela Krasteva

Current trends in electrosurgery include digital algorithms and methods to control the effect of

delivered energy to the patient. Process control of the delivered radio frequency power is used for various tissue treatments. A control unit in the form of digital signal processor (DSP) or field programmable gate array (FGPA) could be used to automatically supply the appropriate power. During a regular electrosurgical procedure i.e. cutting, coagulation or desiccation, electrical sparks are formed as part of the process. Electrosurgical sparks can create undesirable physiological effects like tissue carbonization and burns. This paper shows an experimental method for high-resolution monitoring of the output current waveform in the process of arcing. The recorded signals provide evidence that current waveform high-pass filtering and amplitude spectrum analysis are meaningful approaches for automatic discrimination between conditions of no arcing, low arcing and high arcing.

Keywords – Electrosurgery, Spark detection, Current waveform, Feedback, Digital control

Анализ на формата на изходния ток за контрол върху процеса на искрене в електрохирургически апарати (Виктор Томов, Иво Илиев, Весела Кръстева). Съвременните тенденции в електрохирургията включват използването на цифрови алгоритми и методи за контрол на ефекта от приложената енергия върху пациента. Управлението на процеса за прилагане на радио-честотна мощност се използва при различни процедури за третиране на тъкани. Контролиращо устройство от типа на цифров сигнален процесор или чип с програмируема логика може да се използва за автоматично титриране на необходимата мощност. По време на обикновена електрохирургическа процедура, включваща рязане, коагулация и изсушаване, възникват електрически искри. Тези искри могат да причинят нежелани физиологични ефекти, като овъгляване и изгаряне. Това изследване представя експериментален метод за регистриране на изходния ток с висока резолюция по време на искрене. Записаните сигнали доказват, че високочестотна филтрация и анализ на амплитудния спектър на тока при електрохирургия са подходящи за автоматично разпознаване на различни условия на работа, в т.ч. без искрене, ниски и високи нива на искрене.

Introduction

Basically, electrosurgery is the use of radio frequency (RF) alternating current (AC) to rise the intracellular temperature in order to achieve vaporization or the combination of vaporization with protein dehydration. These surgical effects are also known as cutting and coagulation of the tissue. There are at least two basic mechanisms whereby RF electricity increases cellular and tissue temperature [1]. The majority of heat is an electrical effect of the intracellular and extracellular ions (commonly include electrolytes Na+, Cl-, Ca++, Mg++, etc.) which act as a media for the high frequency current. Therefore, a resistance is encountered by ionic movement as ions collide with other molecules, resulting to heat generation. The other part of the heat is from net

tissue resistance or conductive heat transfer [2], [3]. The central concept in electrosurgery is that any

current passing through the tissue will dissipate energy as heat [4]. Joule’s law commonly approximates the energy quantity in electrosurgical therapy: Q = I2×R×t, where Q is energy (joule, J), I is current (A), R is electrical resistance (Ω) and t is the time interval for current to flow. As the electrodes have relatively small amount of impedance compared to the tissue, the current passing through the closed circuit will dissipate the energy in the form of heat into the tissue resistance. When current carrying electrodes are brought into contact with the tissue, the current spreads out in the tissue volume.

In a monopolar system, the active electrode is relatively small, leading to high current density. On the other side, the dispersive electrode is large enough

“Е+Е”, vol. 53, 1-2, 2018 23

so that the contact electrode-tissue impedance is low, resulting in low current density and low heat dissipation. Assuming adiabatic conditions and no phase difference between the current and voltage (no reactive components), the temperature rise in the tissue volume can be approximated using the function [4]

(1) ctJT σρ/2=Δ where ∆T is the temperature rise, σ is the tissue electrical conductivity, c is specific heat capacity and ρ is the tissue density. Presuming an idealized model of a round electrode positioned in a medium, the power density is inversely proportional and falls off rapidly with the distance to the electrode (2) 422 / raVWv σ= where V is the output voltage, a is the radius of the electrode and r is the distance from the electrode.

One aspect in electrosurgery is to control the temperature and the time of the tissue exposure to this temperature. Applying exact amount of energy to achieve specific thermal effect is vital for patient recovery. Adverse thermal spread should be reduced to minimum. The mechanisms to control the tissue temperature exposure are closed loop systems using feedback parameters, such as temperature sensors, current and voltage waveform recognition, smart impedance sensing or combination between different methods.

The aim of this paper is to present the current trends for advance power control and spark detection. Experiments are performed to observe the process of arcing and its effect on the output current waveform.

Arcing and cutting effects on tissue

Due to the power dissipation, the tissue temperature increases, with highest temperatures right next to the active electrode. Some cytochemical processes occur in a tissue volume experiencing temperature increase up to 45°C. However, those processes are reversible and the tissue cells return to their normal state. In general, temperatures above 45°C cause breakdown in the living structure and degradation in the protein molecule. Depending on the applied time, temperatures between 45°C and 60°C solidify the proteins inside the cells during a process named coagulation. Up to 100°C slow boiling of the intracellular fluid occurs. As a result, several cells link up and form welding effect, which stops bleeding. Around 100°C vaporization of the tissue fluid within

the cell membrane occurs, resulting in an explosion of the cells and forming a steam around the electrode. This process is termed electrotomy or cutting. Further increase of the temperature is causing carbonization of the tissue, leading to electrical impedance increase and lowering the current density [5].

During the cutting procedure, a desiccation close to scalpel electrode is formed. This desiccation pushes away the tissue front so that a gap between the active electrode and the tissue itself is formed. This gap is filled with various vapours during the electrotomy process (Fig.1). The steam channel makes good condition for sparks to break through between the electrode and the desiccated tissue. Due to the high current, whenever a spark is created on a local point, carbonization takes place in a small area around the electrode. Observations show that the degree of carbonization is increased with the increase of the light emissions around the active electrode. Furthermore, the intensity of arcing is highly dependent on the tissue type [6]. Carbonization effect is influenced mostly from the amount of sparks instead of the applied power. In order to minimize the unnecessary carbonization and tissue necrosis, it is assumed that the cut procedure mechanism should include feedback control not only based on the output power but also on the spark rate [6].

Fig.1. Sparks inside the steam channel while cutting.

Current waveform performance under arcing conditions

To generate a spark, electric field is developed reaching a high enough value for creating a medium for breakdown. Typically, a breakdown could be achieved at 3 MV/m for dry air [7].

Fig.2 shows the volt-ampere (V/A) characteristic of an electrosurgical process, which includes arcing conditions. The hysteresis is due to the intrinsic circuit reactance and the thermal misbalance between the

24 “Е+Е”, vol. 53, 1-2, 2018

particles [7]. The current is having small increment till 500 V, while significant positive increment is observed from 500 V to 1000 V. The latter represents conditions where arcing process is formed [7]. We can conclude that during spark formation a rapid current increase is observed overlayed with the sine wave output of the generator. Furthermore, sparks could be detected by the rapid peaks in the current-time waveform [6].

Fig.2. Electrical VA characteristic of the electrosurgical

process [7].

Closed loop control and digital processing while arcing in electrosurgical system

Most of the electrosurgical generators on the market support operator-driven adjustment of the output power. Depending on the specific procedure, the proper power level could be manually selected. Some typical power values are listed in Table 1 [5].

Table 1

Typical surgical procedures using electrosurgical generators

Power Level Procedures Low power 30 W cut 30 W coagulation

Neurosurgery Dermatology, Plastic surgery

Medium power 30 W - 150 W cut 30 W - 70 W coagulation

General surgery Major vascular surgery

High power 150 W cut 70 W coagulation

Transurethral resection Ablative cancer surgery

In order to achieve the required performance, a

digital power control unit could be implemented in the electrosurgical device. Computer-based systems are available for continuous monitoring of the current and voltage levels; therefore, an automatic power adjustment could be implemented to preserve the

surgical effect while working on different types of tissues with various electrical impedances. The design of such system is presented in Fig.3. The output signals (VSENS, ISENS) are monitored and digitized by analog to digital converters (A/D). Furthermore, the computing unit implements a signal processing algorithm for calculation of the optimal power delivery. The result is used to take the decision for increment or decrement of the generator’s output current in order to maintain the set power demand. Moreover, the digital signal processing of the current and voltage data streams could be used to further recognize distinguishable waveform characteristics under specific operating conditions, i.e. while arcing or drying out of tissue [8].

If we consider that the arcing process includes rapid current spikes or high frequency components, a high frequency filter could be implemented to detect this specific condition. A simplified digital high-pass filter in the form of first-difference differentiator of the successive samples x(n), x(n-1) (Eq.3) is appropriate [9]

(3) )1()()( −−= nxnxny .

Fig.3. Simplified block diagram of a closed loop scheme.

More complex method to analyze the voltage and current data streams is to perform discrete Fourier transform (DFT). The DFT of a sample vector X(j) with length n is defined as follows:

(4) ( 2 )( 1)( 1)

1

1 ( )i j kn

n

j

kY X j e

nT

π− − −

=

− =

,

“Е+Е”, vol. 53, 1-2, 2018 25

where k follows the sequence 1 to n, j is the sample from the time domain vector and T is the sampling period [10].

We suggest that the design of an algorithm that analyzes the power spectrum Y could be a powerful solution for the reliable detection of spark conditions.

Experimental setup

A commercial electrosurgical generator is selected for the purpose of this experiment (Bovie’s IDS-310 model). Fig.4 shows simplified block diagram of the measurement setup.

Fig.4. Experimental setup to investigate the current waveform in spark conditions.

The experiment is performed on a simulated patient tissue in the form of non-inductive resistance. To achieve consistent measurements, a 300 Ω resistor is connected between the neutral and the active electrode of the generator. The current waveform is recorded using high speed oscilloscope at various test scenarios described below. The sampling rate of the oscilloscope is 25 MS/s (the period between the samples is 40 ns).

Results

The experiments are performed while the electrosurgical generator is set to output power (20 W), cut mode (492 kHz sinewave), output load (300 Ω). The waveform is recorded in the following conditions: • No arcing – both active and neutral electrodes are

firmly connected to the non-inductive load. The snapshot of the waveform is shown in Fig.5.

• Low arcing – a metal plate is connected at one end of the load. Using the monopolar pencil with disposable electrode, the hand-piece is approached to the metal plate. The active electrode is moved and kept at distance from the metal plate, such that small arcing film is created. At this moment a snapshot of the waveform is recorded (Fig.6).

• High arcing – the same setup as in low arcing is used, but the distance between the active electrode

and the metal plate is increased. Furthermore, more aggressive bursts of arcs are created. Snapshot of the waveform is recorded (Fig.7).

Fig.5. Current waveform, no arcing, 20 W power setting,

300 Ω output load.

Fig.6. Current waveform, low arcing, 20 W power setting, 300 Ω output load.

Fig.7. Current waveform, high arcing, 20 W power setting, 300 Ω output load.

26 “Е+Е”, vol. 53, 1-2, 2018

All recorded current signals in Fig.5-Fig.7 are further passed through the first-difference filter (Eq.3), and the resulting high-pass components are illustrated in Fig.8-Fog.10, respectively.

Fig.8. High-pass components of the current waveform in Fig.5, no arcing, 20 W power setting, 300 Ω output load.

Fig.9. High-pass components of the current waveform in

Fig.6, low arcing, 20 W power setting, 300 Ω output load.

Fig.10. High-pass components of the current waveform in Fig.7, high arcing, 20 W power setting, 300 Ω output load.

At the next step, the DFT spectrum of all recorded current signals in Fig.5-7 is calculated and illustrated in Fig.11-13, respectively. The amplitude spectrum is represented up to the Nyquist frequency, equal to 12.5 MHz (half of the sampling rate).

Fig.11. Amplitude spectrum of the current waveform in

Fig.5.


Fig.6.


Fig.7.

“Е+Е”, vol. 53, 1-2, 2018 27

Discussion and Conclusions

This work shows that sparks generate significant complexity of the current waveform. In particular, the arcing process adds high frequency components to the waveform. As shown in Fig.8-10, a simple first-difference filter could be used to enhance the high frequency components, where rapid current bursts during sparking conditions are recognized. Comparing Fig.9 (low sparks) to Fig.10 (high sparks), we see evidence that the more the arcing process is, the higher peaks in the high-pass filter output appear. For reference, we do not notice any peak bursts in the high-frequency current components without arcing.

Considering the DFT of the current waveforms (Fig.11-13), we can confirm that the higher sparks induce more high frequency components. In the first case of no arcing (Fig.11), the spectrum is formed mostly of the base harmonic (492 kHz) together with small amount of third harmonic (1476 kHz), which represents the output characteristic of the generator. In case of low arcing (Fig.12), additional high frequency components >2 MHz are observed. In the case of high arcing (Fig.13), the spectrum includes more frequency harmonics together with lower amplitude of the base frequency compared to low arcing in Fig.12.

To control the amount of arcing, current waveform analysis could be implemented in the form of: • High frequency filter together with peak detection

to measure the rate and periodicity of peak events. • Crest factor (CF) measurement - CF is equal to

the ratio of the peak value to the root mean square (RMS) value. Typically, the crest factor of a sinewave is 1.41. Rapid bursts in the sinewave current will lead to increase of the CF value.

• The increase of the current during a spark is introducing high frequency components. Low frequency moving average filter after a high-pass one could be implemented to detect the amount of arcing. Additionally, DFT could be introduced to detect the high order harmonics.

Advanced electrosurgical devices (Fig.3) include mechanism and sensors to measure the current waveform in a real-time process. A computing unit in the form of DSP or FPGA is appropriate to implement the above stated methods for spark control.

Further developments could include time and frequency domain signal processing of the measured waveforms for the design of loop algorithms for spark control. Moreover, the voltage and current waveforms could be synchronously proccessed and advanced measurements could be implemented in the spark detection algorithm, such as complex impedance monitoring and measurements on VA characteristics.

REFERENCES

[1] Feldman, L.S. et al. eds. The SAGES manual on the fundamental use of surgical energy (FUSE). Springer Science+Business Media, LLC 2012.

[2] Vilos, G., C. Rajakumar. Electrosurgical generators and monopolar and bipolar electrosurgery. J Minim Invasive Gynecol, 2013;20(3):279–287.

[3] Pfiedler Enterprises. Electrosurgery. An online continuing education activity, Denver, 2016.

[4] Frich, L. Radiofrequency ablation of liver tumors, an experimental and clinical study. University of Oslo, Faculty of Medicine, 2007.

[5] Eggleston, J.L., W.W. Von Maltzahn. Electrosurgical devices. CRC Press LLC, 2000.

[6] Wurzer, H., R. Maeckel, J. Lademann, H. Audring, D. Liess. A spark counter as a control unit of a radio frequency surgery device. IEEE Transactions on Biomedical Engineering, Vol. 44, 1997, pp. 831 – 838.

[7] Schneider, B., E. Dias, P.J. Abatti. How can electrosurgical sparks generate undesirable effects?. First IEEE Latin American Symposium on Circuits and Systems (LASCAS), 2010.

[8] Tomov. V. Modern Advances in energy based electrosurgical devices. Electronica, 2017, pp. 81–84.

[9] Lyons, R. A differentiator with a difference. Blog, November 3, 2007, https://www.dsprelated.com/ showarticle/35.php

[10] Matlab documentation, Fast Fourier transform function (fft) https://www.mathworks.com/help/matlab/ref/fft.html.

Eng. Viktor T. Tomov – R&D Engineer at Bovie-Bulgaria, part of Bovie Medical Inc. Currently doing PhD thesis at Technical University of Sofia, Department of Electronics. Interested in design and development of electrosurgical apparatus and more specific control algorithms for specific surgical effects. Furthermore, cold plasma applications for treating tissue using inert gases and high voltage generators.

tel.:+359 889 534036 е-mail: [email protected]

Prof. Ivo Ts. Iliev, DSc – Head of Lab. of Biomedical Engineering, Department of Electronics at the TU-Sofia. Areas of scientific research – Medical instrumentation and Biomedical signal processing.

Vessela Krasteva, PhD – Assoc. Prof. at Institute of Biophysics and Biomedical Engineering – Bulgarian Academy of Sciences. Her scientific achievements are related to development of methods, algorithms and software applications in the field of biomedical signal processing and electrical therapy, with applications to ECG instrumentation and defibrillation.


28 “Е+Е”, vol. 53, 1-2, 2018

Study of the factors influencing power consumption of FPGA-based designs

Galia Marinova, Zdravka Tchobanova

The paper presents results from the study of power consumption of basic circuit designs on

FPGA with different VHDL descriptions, different elaborated and synthesized structures and realized on different devices. The study is performed for 4 different VHDL descriptions (behavioral and structural) of 4-bit comparator and realizations on Xilinx FPGA circuits from the families Artix-7, Zynq 7000 all programmable System-on-a-chip (SoC) or on the Xilinx board Zedboard. The Xilinx software Vivado 2014 is used and the function Power report is implemented for power consumption study. Power consumption is sensitive to factors enumerated. For the 4-bit comparator studied it goes to 29% increase of dynamic power for structural VHDL description and it shows that the low-power design on FPGA should start with the low power design of very basic structures as combinatorial logic circuits.

Keywords – 4-bit comparator, FPGA, Power consumption, VHDL code

Изследване на факторите влияещи на консумацията на проекти, реализирани върху програмируеми схеми FPGA (Галя Маринова, Здравка Чобанова). Статията представя резултати от изследването на консумираната мощност на схемни проекти, реализирани върху програмируеми схеми FPGA, основани на различни описания на VHDL кода (поведенчески и структурни), различни логически и синтезирани структури, и реализирани върху различни устройства. Изследването е направено за 4 различни описания на VHDL кода на 4-битов компаратор и последващата им реализация върху Xilinx FPGA схеми от фамилиите Artix-7, Zynq 7000 напълно програмируеми системи-върху чип (SoC) или върху платката на Xilinx Zedboard. Използван е софтуерът за цифрово проектиране Vivado 2014 на Xilinx, а резултатите за изследваната мощност на всяка проектна реализация са получени с опцията Power report. Изследването показва, че консумираната мощност зависи от изброените по-горе фактори. При структурно описание на VHDL на 4-битовия компаратор се получава до 29% нарастване на изразходваната динамична мощност в сравнение с поведенческото описание, следователно проектирането върху FPGA за ниска консумация би следвало да започне с проектирането за ниска консумация на базисните схемни структури, например комбинационните логически схеми.

1. Introduction

The paper presents the results of a research, part of a project in Technical University-Sofia on design technics for green telecommunications. An overview on green communications is presented in [1] and some results of power consumption power estimation on FPGA and USRP-based platforms are presented in [2]. As noticed in [5] FPGA are 12 times less power efficient than ASICs because of the large number of transistors per logic function, necessary for programming. Static power in CMOS circuits is the power consumed when the circuit is not active, dynamic power is the power consumed when the logic

toggles. Dynamic power is predominant in CMOS based devices - CPLD, FPGA, digital ASICs and it's a sum of switching power due to charging and discharging of output capacitance and short circuit power due to non-zero rise/fall times. It can be calculated as [4]:

(1) fVQVCP circuitshortdyn ..21 2

+= − ,

where Pdyn - dynamic power C – load capacitance, V-power supply voltage, f - switching frequency, Qshort-

circuit - short-circuit charge. The average power consumption in an FPGA can

be calculated as:

“Е+Е”, vol. 53, 1-2, 2018 29

(2) 212

1VfCP i

n

i iavg == ,

where Pavg - average value of dynamic power dissipation, Ci – capacitance of the net i, V-power supply voltage, fi - average switching frequency, n- number of nets.

Recently different studies are performed in order to model and reduce power consumption in FPGAs, as shown in the overview from [3]. Dynamic power consumption is architecture dependent [6], data dependent, hardware description language (HDL) code dependent [5]. Different techniques are proposed to reduce dynamic power on clock scheme, logic power, RAM power, I/O power by optimal selection of adder and multiplier blocks, counters, FSMs, general glitch reduction techniques on logic, logic partitioning rearranging, etc.

Authors of [5] study 2 HDL codes of a 4-bit unsigned up counter with asynchronous clear and clock enable on Xilinx Virtex-6 FPGA - the first code maps the clock enable signal to LUTs and the second maps it to the control ports and they obtain 6% power consumption reduction in the first design. The study in the current paper is focused on the influence of VHDL codes of combinatorial circuits’ designs, implemented on FPGAs, on their power consumption and the impact of FPGA device on the power consumption of the design. The design studied is a 4-bit comparator and the simulator used is Vivado 2014. It is selected because it is realized with XOR circuits, which have significant power consumption [7]. The FPGA implementations and devices are studied, because they are widely used for communication systems [8]. The size of 4-bit was selected for comparison of behavioral and structural VHDL descriptions.

2. Study of the VHDL description influence on power consumption of FPGA-based designs

Four different VHDL descriptions of a 4-bit

comparator are studied for power consumption in this section. The entities (ekv1, ekv2, ekv3, ekv4) of the 4-bit comparator, described in VHDL are identical for the 4 different architectures as shown on Fig.1. entity ekv1 is Port ( a : in STD_LOGIC_VECTOR (0 to 3); b : in STD_LOGIC_VECTOR (0 to 3); a_eq_b : out bit); end ekv1;

Fig.1. Entity of the 4-bit comparator.

Table 1 shows the 4 different VHDL codes of the 4 different architectures for the 4-bit comparator, implemented on ZedBoard with circuit Zynq™-7000 SoC XC7Z020-CLG484-1 [10]. The behavioural simulation of the designs is presented on Fig. 2.

Fig.2. Simulation of 4-bit comparator in Vivado 2014.

For each description elaborated and synthesized designs are obtained with the options:

Implemented design > Report power in VIVADO 2014, as shown in Table 1. The numbers of cells and nets differ for each description: (1 Cell, 9 Nets), (8 Cells, 16 Nets), (1 Cell, 9 Nets), (5 Cells, 13 Nets) for the elaborated designs and (6 Cells, 4 Nets), (11 Cells, 19 Nets), (11 Cells, 19 Nets), (11 Cells, 19 Nets) for the synthesized designs. The basic elements of the synthesized designs are 1 LUT3 and 1 LUT6 and different buffers. The first 3 VHDL descriptions are behavioral and the forth description is structural. Behavioral descriptions in VHDL are based on logical equations and truth tables and structural descriptions are based on schematic description with components and connections between them. Table 3 presents the on-chip power consumption of designs from Table 1. The on-chip power values for the first 3 descriptions are very close and they have similar structure for total power, dynamic power, static power, power dissipation in signals, logic and I/Os. The structural description leads to 29% larger total on-chip power consumption and different repartition of the on-chip power 71% dynamic power/ 29% static power compared to 63% dynamic power/ 37% static power in the 3 designs with behavioral architectures. Power dissipation in Logics stays almost invariant in all 4 designs - with a value of 0.03-0.04 W or 2% of the total on-chip power for behavioral descriptions and 1% for structural description. I/O power increases with 13% in the design with structural description. The largest difference is in the values of power for signals. The increase for structural description is 4.29 times the power for behavioral descriptions. The static power consumption stays almost invariant - 0.122-0.123 W and it's 37% of total on-chip power for

30 “Е+Е”, vol. 53, 1-2, 2018

behavioral descriptions and 29% of total on-chip power for structural description. This study shows that the VHDL description of a design - behavioral or structural, influences the dynamic power consumption

of the corresponding design and almost not the static power consumption. The most influenced elements of power are signals and I/Os.

Table 1 Power consumption of a 4-bit comparator project on XILINX FPGA,

designed from different VHDL descriptions

VHDL code Elaborated design Synthesized design On-chip Power architecture Behavioral of ekv1 is begin process (a,b)

begin if a=b then a_eq_b<='1'; else a_eq_b<='0'; end if; end process;

end Behavioral;

1 Cell 9 Nets

6 Cells 4 Nets

architecture Behavioral of ekv2 is begin a_eq_b<= not((a(0)xor b(0))or (a(1)xor b(1))or (a(2)xor b(2))or (a(3)xor b(3))); end Behavioral;

8 Cells 16 Nets

11 Cells 19 Nets

architecture Behavioral of ekv3 is begin a_eq_b<= '1' when (a=b) else '0'; end Behavioral;

1 Cell 9 Nets

11 Cells 19 Nets

architecture Structural of ekv4 is component xor2 port (a,b:in std_logic; q: out std_logic); end component; component nor4 port (a,b,c,d: in std_logic; qn:out std_logic); end component; signal c:STD_LOGIC_VECTOR (0 to 3); begin x0: xor2 port map(a(0),b(0),c(0)); x1: xor2 port map(a(1),b(1),c(1)); x2: xor2 port map(a(2),b(2),c(2)); x3: xor2 port map(a(3),b(3),c(3)); a1: nor4 port map (c(0),c(1),c(2),c(3), a_eq_b); end Structural;

5 Cells 13 Nets

11 Cells 19 Nets

“Е+Е”, vol. 53, 1-2, 2018 31

Table 2 Power consumption estimation of a design implemented on different XILINX FPGAs

FPGA Synthesized design On-chip power

xa7a35tcpg236-2I (active), Family Artix-7

xc7a35tcpg236-3 (active) Family Artix-7

xc7z010clg225-3 (active), Family Zynq 7000 all programmable SoC

ZedBoard with circuit Zynq™-7000 SoC XC7Z020-CLG484-1

Table 3

On-chip Power consumption in [W] of designs from Table 1

Table 4 On-chip Power consumption in [W] of designs in Table 2

FPGA Total On-chip

Power

Static Power

Dynamic Power

Power in

Signals

Power in

Logic

Power in I/O

xa7a35tcpg236-2I

0.277 0.071 0.206 0.022 0.003 0.181

xc7a35tcpg236-3

0.37 0.071 0.299 0.090 0.004 0.205

xc7z010clg225-3

0.37 0.071 0.299 0.090 0.004 0.205

ZedBoard 0.328 0.122 0.206 0.021 0.03 0.181

VHDL Code

Total On-chip Power

Static Power

Dynamic Power

Power in Signals

Power in Logic

Power in I/O

Ekv1 0.328 0.122 0.206 0.021 0.003 0.181 Ekv2 0.329 0.122 0.207 0.021 0.004 0.181 Ekv4 0.328 0.122 0.205 0.021 0.003 0.181 Ekv4 0.423 0.123 0.299 0.090 0.004 0.205

32 “Е+Е”, vol. 53, 1-2, 2018

The power consumption of logic is slightly affected by the VHDL description. The maximal delay from input to output ports in the 3 designs with behavioral description is 7.279 ns and for the design with structural description, the value is slightly higher 7.411 ns. The results show that at early stage of design, it is important to select a low power VHDL description for the basic combinatorial circuits in order to help the low power design of a more complicated system on FPGA.

3. Study on the influence of device selection on the power consumption of FPGA-based design

In this section the influence of device selected for a design is studied. The 4-bit comparator ekv1 with the first behavioural description from Table 1 is implemented on 4 different Xilinx FPGAs:2 from the family Artix-7 [9]: xa7a35tcpg236-2I and xc7a35tcpg236-3 and 2 from the family Zynq™-7000 - the first one is xc7z010clg225-3 and the second one is XC7Z020-CLG484-1 on Zedboard [10].

The VHDL description of the architecture used is:

architecture Behavioral of ekv1 is begin

comp: process (a,b) begin if a=b then a_eq_b<='1'; else a_eq_b<='0'; end if;

end process comp; end Behavioral;

Table 2 presents the synthesized designs for each

device and the data for powers consumption. Table 4 presents the values for on-chip power consumption of designs in Table 2.

The total on-chip power is lower for the device xa7a35tcpg236-2I - 0.277 W and it's higher for xc7a35tcpg236-3 and xc7z010clg225-3 - 0.37 W, which results in 34% difference.

All 4 implementations have different repartitions between dynamic and static power:74% - 26% for xa7a35tcpg236-2I , 81% - 19% for xc7a35tcpg236-3 and xc7z010clg225-3, 63%-37% for Zedboard. Static power for the devices xa7a35tcpg236-2I, xc7a35tcpg236-3 and xc7z010clg225-3 is identical - 0.071 W, but for Zedboard it's 0.122 W, which is 72% larger.

Dynamic power in xa7a35tcpg236-2I and Zedboard is 0.206 W and in xc7a35tcpg236-3 and xc7z010clg225-3 it's 0.299 W, which is 45% larger.

Dynamic power in Logic is almost identical in 4 devices 0.03-0.4 W (1-2%).

Power in I/Os for xa7a35tcpg236-2I and Zedboard is 0.181 W and in xc7a35tcpg236-3 and xc7z010clg225-3 it's 0.205 W, which represents 13% increase.

Power consumption in signals has the most significant increase - for xc7a35tcpg236-3 and xc7z010clg225-3 the increase is 4.29 times, compared to the power consumption in signals for xa7a35tcpg236-2I and Zedboard implementations.

These data show that the selection of the device for implementation of a design has a serious impact on its power consumption, both dynamic and static, and it has to be considered. As in previous section, the largest differences are in power consumption for signals and there are very slight or null difference for power consumption in logic.

4. Conclusion

The study performed on the impact of different VHDL description (and corresponding elaborated and synthesized schematics) of a circuit design, show that the description might strongly increase the power consumption. The design of a 4-bit comparator with structural description, implemented on Zedboard, shows 29% larger consumption compared to the 3 behavioral descriptions of the same circuit. Static power consumption of the 4-bit comparator is almost not affected by the VHDL description. Structural description is very intuitive for the designer but it’s not optimized in terms of complexity in implementation. This explains the higher power consumption in structurally described designs. Structural description leads to slightly higher maximal port to port delay.

The repartition of power consumption depends both from the VHDL description and from the device on which the 4-bit comparator is implemented. The power consumption in signals is the most sensitive to different VHDL descriptions and to different devices. It goes to 4.29 times of difference. Dynamic power consumption for logic seems the least sensitive to VHDL description code and to device selection. Power in I/Os varies about 13% from the VHDL code description and device selection.

The results from this study show that even for combinatorial circuits the type of VHDL code description has important influence on power consumption. Structural description might increase the power consumption. So, when a more complicated design is planned, the basic structures should be designed initially for low power, in order to achieve lower power consumption of the final complex design. On the other side the type of device selected for a

“Е+Е”, vol. 53, 1-2, 2018 33

design is also influencing the power consumption, so it should be considered at the early stage of the design, when basic structures are described and tested.

Further research is foreseen on clock-driven designs as random bit and number generators.

This paper is an extension of work originally

reported in the National forum with international participation ELECTRONICA, Sofia, Bulgaria, 2017.

Acknowledgements

The research described in the paper is partly supported by Project 162PD0020-07 in Technical University – Sofia.

REFERENCES

[1] Tchobanova, Z., G. Marinova. Telecommunication system for green economic – a survey. ХХIV TELECOM’2016, 27-28 October, NSTC, Sofia, Bulgaria, pp.177-186

[2] Marinova, G, Z. Tchobanova. Circuit design for green communications – methods, tools and examples., Proc. of the 5th UBT Annual Int. Conf. on Business, Technology and Innovation, Chapter: Computer Science and Communication Engineering, Durres, Albania, 28-39 October 2016, pp.27-37

[3] Grover, N., M.K. Soni. Reduction of power consumption in FPGAs – An Overview. Int. J. Information Engineering and Electronic Business, 2012, 5, pp.50-69.

[4] Ibro, M., L. Karcanaj. Models for reducing power consumption in CPLD and FPGA devices. Proc. of the 5th UBT Annual Int. Conf. on Business, Technology and Innovation, Chapter: Mechatronics, Sciences in Energy Efficiency Engineering, System Engineering and Robotics, Durres, Albania, 28-39 October 2016, pp.12-19.

[5] Pandey, B., M. Pattanaik. Low power VLSI circuit design with efficient HDL Coding. Communication Systems and Network Technologies (CSNT), 2013 International Conference on, 6-8 April 2013, pp. 698-705.

[6] Li, F., D. Chen, L. He, J. Cong. Architecture evaluation for power-efficient FPGAs. FPGA’03, February 23-25, 2003, Monterey, California, USA, pp.175-184.

[7] Ye, Y., K. Roy, R. Drechsler. Power consumption in XOR-based circuits. IEEE Asia and South Pacific Design Automation Conf. ASP-DAC'99, Hong Kong, 1999, pp. 299-302.

[8] Masselos, K., N.S. Voros. Implementation of wireless communications systems on FPGA-based platforms. EURASIP Journal on Embedded Systems 2007(1):1-1, January 2007, Hindawi Publishing Corporation, doi:10.1155/2007/12192, Available from: https://www.researchgate.net/publication/234805823_Implementation_of_Wireless_Communications_Systems_on_FPGA-Based_Platforms [accessed Mar. 11 2018].

[9] XILINX ARTIX-7 FPGAS: New performance and bandwidth standards for power-limited, cost-sensitive markets, https://www.xilinx.com/support/documentation/product-briefs/artix7-product-brief.pdf

[10] ZedBoard (Zynq™ Evaluation and Development) Hardware User’s Guide, http://zedboard.org/sites/default/files/documentations/ZedBoard_HW_UG_v2_2.pdf

Assoc. Prof. Dr. Galia Il. Marindva graduated as engineer with Master degree in electronics in 1988 in the Faculty of electronics in Technical University-Sofia. She received a Ph.D. degree in 1994 in the Faculty of electronics in TUS. Since 2011 she’s associate professor in the Faculty of Telecommunications in TUS. She did one year post-doctoral research in CNAM-Paris, France in 1999/2000. She is author and co-author of more than 80 scientific papers, mainly in the area of computer-aided design in electronics and telecommunications. She got the Certificate of merit at the World Congress on Engineering in London, UK in 2007 and recently a Best paper award at the Twelfth Advanced International Conference on Tele-communications, AICT 2016, May 22 - 26, 2016 - Valencia, Spain. Galia Marinova is coordinator of CEEPUS network project and has experience with several Erasmus+ projects.

tel.: +359 (2) 965-3188 е-mail: [email protected]

Eng. Zdravka N. Tchobanova - received the M.E. degree in communication engineering in 1986 from Technical University of Sofia. She is currently Ph.D. students in the Faculty of Telecommunications in Technical University - Sofia. Her research interests are spectrum sensing in cognitive radio, power consumption in software radio systems, algorithms and architectures for communications on platforms with FPGA and USRP.

tel.: +359 (0)888 693713 е-mail: [email protected]


34 “Е+Е”, vol. 53, 1-2, 2018

TELECOMMUNICATIONS SCIENCE

Multi-access edge service for application-initiated offload

Evelina Pencheva, Denitsa Kireva, Ivaylo Atanasov

The growth in services for smart devices, such as video, music, social networking, gaming and other

interactive applications, that become ubiquitously penetrated, sets requirements for high bandwidth and low la-tency, which the traditional centralized network architecture cannot face due to heavy burdens on the backhaul links and long delays. Multi-access Edge Computing (MEC) is an edge paradigm which deploys cloud services in the radio access network. The vicinity to the end users allows timely reaction to dynamic changes in radio conditions. Offloading is a promising technique for reducing the amount of data being carried on the cellular bands to alternative wireless ones. With existing solutions, it is the user who initiates offloading. In this paper, we propose a new mobile edge service that allows mobile edge applications to initiate offloading. The offloading trigger may be congestion in the radio access network, quality of service experienced by the user, user location, etc. The service is described by use cases, a data model and application programming interfaces. The data mod-el represents the resources as a tree structure, provides applications with data-mediation functions as far as it is a meaningful way for addressing resources. As a proof of concept, models representing the offloading state as seen by the application, serving and target access networks are proposed and formally verified.

Keywords – Edge computing, Proximity-based service, Application programing interfaces, Finite state machines

Услуга за множествен достъп за приложения, иницииращи разтоварване (Евелина Пенчева, Деница Кирева, Ивайло Атанасов). Нарастването на броя на услугите за интелигентни устройства като видео, музика, социални мрежи, игри и други интерактивни приложения, които се разгръщат пов-семестно, поставя изисквания за широка честотна лента и ниска латентност, които традиционната централизирана мрежова архитектура не може да предложи поради натоварването на връзките в опорната мрежа и големите закъснения. Multi-access Edge Computing (MEC) е парадигма за края на мрежата, която разгръща облачните услуги в мрежата за радиовръзка. Близостта до крайните пот-ребители позволява своевременна реакция при динамични промени в радиоусловията. Разтоварването е обещаваща техника за намаляване на количеството данни, пренасяни от клетъчните радиомрежи към алтернативни безжични такива. При съществуващите решения потребителят започва разтоварване-то. В тази статия предлагаме нова услуга в мрежата за радиодостъп, която позволява на мобилните приложения да инициират разтоварване. Причина за разтоварването може да бъде претоварване в мрежата за радиодостъп, качеството на обслужване, предоставено от потребителя, местоположе-нието му и т.н. Услугата е описана посредством случаи на използване, модел на данни и приложни програмни интерфейси. Моделът за данни представя ресурсите като дървовидна структура и предос-тавя на приложенията с функции за медиация на данни, доколкото това е начин за адресиране на ре-сурсите. Като доказателство на концепцията са предложени и формално проверени модели, предста-вящи състоянието на разтоварване, от гледна точка на приложението, обслужващата и целевата мрежа за достъп.

Introduction

The ubiquitous penetration of smart devices with embedded capabilities for multi-radio technologies operation and the exposure of attractive applications like video, streaming, gaming, social networking etc. become a challenge for traditional telecom operators which have deployed centralized architectures. These architectures feature long latency and heavy load on the backhaul links. Therefore, new architectures that bring the network function control close to the edge of

the mobile network are needed. Multi-access Edge Computing (MEC) is a relatively

new paradigm that brings cloud capabilities into the radio access network [1]. The vicinity to the end users enables applications with low latency and high band-width requirements, which may be provided with real time radio network information and location awareness [2].

MEC may be useful in migrating cellular data traf-fic to supplementary networks using different com-

“Е+Е”, vol. 53, 1-2, 2018 35

munication technologies such as Wireless Access Network (WLAN) and thus mitigating the radio access network overloading [3], [4].

To provide seamless roaming and service continuity there is a need to combine various wireless access tech-nologies and provide reliable intersystem handover strategies [5]. Related works on intersystem handover are focused on performance optimization [6], [7], [8] and algorithms for handover decision [9], [10], [11].

With the current solutions, it is the User Equipment (UE) that initiates offloading. The network provides access network discovery and selection information which reflects the telecom operator preferences for access prioritization, but the final choice is delegated to the user.

In this paper, we propose a new mobile edge ser-vice which allows mobile edge applications to initiate offloading based on application specific policies. The triggers for application-initiated offloading may be real-time information about access network conges-tion, user location, quality of service experienced by the user, requested data speeds, call drops, and even information about available user credit.

Following the adopted RESTful style for definition of mobile edge service interfaces, we model resources as tree structure which provides great flexibility of addressing resources and creation of new resources. We also discuss actual interface definition by provid-ing HTTP methods that may be used for resource ma-nipulation. To prove our concept, we describe the offloading state models which represent the applica-tion’s, serving and target access networks’ views, and show that these models expose equivalent behavior.

The paper is structured as follows. Next section provides an informative description of the new mobile edge computing service, illustrating the service func-tionality by a use case. Then a data model that pro-vides data-mediation functions is described and some examples of mobile edge service interfaces are pro-vided. Before the conclusion, the offloading models as seen by the application, serving cellular access net-work, and target WLAN, are described and formally verified. The conclusion summarizes the author con-tribution.

Informative service description

The proposed mobile edge service for WLAN of-floading allows applications running at the edge of the network to control mobile data offloading i.e. to trig-ger usage of WLAN technology for delivering data targeted originally for cellular network.

There are several different technologies that fall in-to the WLAN category. An existing industry standard is IEEE 802.11b operating in the 2.4 GHz ISM band.

A new entrant for this same band is Bluetooth. New technologies such as IEEE 802.11a and ETSI BRAN Hiperlan2 are being developed for 5GHz bands. It is recognized that WLANs are, and will continue to be deployed by independent (i.e. non 3GPP) operators and that these WLANs may or may not be interworked with a 3GPP system. Further, these WLANs may overlap with WLANs that are interworked with 3GPP systems, and they may also overlap each other. These situations create multiple permutations of coverage areas and service states which will need to be carefully understood and managed.

Mobility is a mechanism of switching between 3GPP (e.g. LTE) and non-3GPP (e.g, WLAN) net-works. For a subscriber, the motivation to switch the used communication channel from LTE to WLAN during connection is to get access to wider bandwidth and probably in lower cost or free. For mobile network operators it would help to reduce the load on the LTE network by offloading the subscriber to WLAN net-work. The network operator may not be able to charge for WLAN network usage as much as for LTE net-work, but they may be able to get some gain from load balancing and still keep some portions of the money from the mobile user by directing to switch to the WLAN network serviced by the mobile network oper-ator (not free WLAN).

With existing standards, it is the user that may ini-tiate WLAN offloading [12]. It is the network that provides information about access network discovery and selection, and thus controls offloading between 3GPP and non-3GPP access networks. The purposes of access network discovery and selection function is to assist UE to discover access networks in their vicin-ity and to provide policies to prioritize and manage connections to available networks, but the final deci-sion on the choice of access network is up to the user. The UE periodically performs WLAN scanning and when a WLAN network is found, an offloading pro-cedure is initiated where the user is prompted to select what network technology to be used.

The research novelty is that with the proposed ap-proach, it is an authorized mobile edge application that may initiate WLAN offloading based on specific ap-plication defined policy. For example, the application may trigger WLAN offloading in case of radio access network congestion or based on user location. Another reason may be quality of service (QoS) experienced by the user or requested uplink and/or downlink data speeds. An authorized mobile edge application may use MEC Location API to receive information about UE location, or Open Mobile Alliance (OMA) diagnostics and trap monitoring mechanism to acquire information

36 “Е+Е”, vol. 53, 1-2, 2018

about available access networks and QoS experienced by the UE [13]. The application may be hosted by net-work operator or by 3rd party. The benefit for deploy-ment the application driven offload functionality as a MEC service is more timely reaction on dynamic changes in radio network conditions. Moving the WLAN offloading trigger logic to 3rd party application enables creation and deployment of attractive applica-tions that implement different offloading policies.

There are two methods for mobility: network initi-ated, and UE initiated. The proposed approach for application driven WLAN offloading is based on net-work initiated mobility.

There exist several 3GPP-WLAN interworking scenarios. The goal of WLAN offloading scenario is to provide seamless service continuity between the ac-cess technologies. Seamless service continuity means minimizing aspects such as data loss and break time during the switch between access technologies.

Use case: Being a 3GPP subscriber, Alice possess-es a VoIP capable, i.e. multimedia, terminal device. She prefers to use WLAN when making her multime-dia calls, because most of the time she communicates happens to be at places where WLAN coverage is available. Still, being mobile, it happens to leave the WLAN coverage while communicating, and therefore it's better for her all ongoing sessions to be kept intact, or the eventual interruption to be as unnoticeable as possible. Alice achieves that by obtaining a WLAN card for her equipment, and thus, switching the 3GPP and WLAN accesses, whenever it's necessary, assures the session continuity.

Fig.1 shows the overall procedure for the case where UE starts communication from LTE and a mo-bile edge application initiates a switch to WLAN (trusted non-3GPP).

The case assumes that UE is connected to LTE be-fore the switch (handover) and not connected in WLAN. If UE is already connected both to LTE and WLAN before this handover, it will skip the steps related to WLAN registration and directly jump to L3 configuration.

The UE has established tunnel for user data to the core network. Following specific policy the applica-tion may initiate WLAN offloading. In order to pro-vide seamless service continuity, this results in inter-system handover from 3GPP serving access network to trusted target WLAN. The serving access network initiates RRC (Radio Resource Control) connection reconfiguration procedure.

The UE in turn initiates WLAN registration which includes EAP (Extensible Authentication Protocol) authentication and WLAN tunnel establishment. L3

Configuration procedure is aimed at acquiring of IPv4 address and/or IPv6 prefix for autoconfiguration. Be-ing configured, the UE may switch to WLAN access. The core network initiates bearer deactivation proce-dure followed by RRC connection release, which in-cludes the release of the established radio bearers as well as all radio resources. The mobile edge applica-tion is notified about the successful offloading to the target WLAN.

Fig.1. Mobile edge application initiated handover from

3GPP access to trusted WLAN.

Data model

REST is an architectural style for distributed sys-tems which features loose coupling of components and stateless interactions. In REST, a distributed ap-plication is composed of resources. A resource can be addressing using HTTP Uniform Resource Identifier

UE Non-3GPP Access

RRC Connection Reconfiguration

3GPP Access Core

network

Established user data

Application

MEC platform

Offload Request

EAP Identity Request

WLAN Registration Request

EAP Identity Response

EAP Challenge Request

EAP Challenge Response

AAA (EAP Identity Request)

AAA (EAP Challenge Request)

AAA (EAP Challenge Response)

AAA (EAP Success)

WLAN Registration Response

Create Session Request

Create Session Response

L3 Configuration

L3 Configuration Complete

RRC Connection EPS Bearer Release

Offload Response

Established user data

“Е+Е”, vol. 53, 1-2, 2018 37

(URI). It has a particular state and may be manipulated by four basic interactions: CREATE, READ, UP-DATE and DELETE. The REST operations can be mapped onto HTTP methods: CREATE is mapped onto POST, READ onto GET, UPDATE onto PUT, and DELETE onto DELETE.

The data model of the proposed mobile edge ser-vice represents the resources as a tree structure. This provides applications with data-mediation functions and it is a meaningful way for addressing resources. The tree structure of resources is illustrated in Fig.2.

Fig.2 Resources representing access network area.

The AccessNetworkArea resource is a placeholder for one or more access network types available for the UE. For the purpose of interworking between 3GPP

access networks and WLAN access networks it is assumed that the available access network types are represented by 3GPPAccessNetworkType and WLANAccessNetworkType resources. The tree struc-ture may be extended in case of offloading to other access networks like 3GPP2 and WiMAX. Each of 3GPPAccessNetworkType and WLANAccess-NetworkType resources acts as a placeholder for one or more access networks of the respective type. Each available access network is represented by a resource, e.g. a WLANAccessNetwork resource contains infor-mation about a WLAN access network available for the UE. The type of 3GPP access network may be E-UTRAN (LTE and LTE Advanced), GERAN (radio access network of GSM/GPRS) and UTRAN (radio access networks of UMTS-based technology).

Each 3GPP access network has attributes that de-scribe the public land mobile network (PLMN) code and cell global identity. Each WLAN access network has attributes that describe the wireless network name. An extended service set (ESS) is one or more inter-connected basic service sets (BSSs) and their associated LANs. Each BSS consists of a single access point (AP) together with all wireless client devices creating a local or enterprise 802.11 wireless LAN (WLAN). The Ex-tended Service Set Identification (ESSID) is one of two types of Service Set Identification (SSID). In an ad hoc wireless network with no access points, the Basic Ser-vice Set Identification (BSSID) is used. In an infrastruc-ture wireless network that includes an access point, the ESSID is used, but may still be referred to as SSID.

Common attributes for all access networks are ConnectivityStatus and ConnectivityActionStatus. The ConnectivityStatus attribute shows the UE con-nectivity status with respect to the particular access network (if it is true, then UE is connected, if false – the UE is disconnected). The Connectivity-ActionStatus attributes show the status of the connec-tivity action (including a progress indicator, a final state and a reminder of the requested action). The ConnectivityAction resource contains sub-resources indicating the handover action which to allow and disable connectivity. The connectivityEnable attribute initiates handover to the target access network and the connectivityDisable attribute initiates the release of radio resources in the serving access network. The execution of an action is triggered by the UPDATE of the corresponding action type.

Following the ETSI way of resources addressing, all resource URIs of the proposed mobile edge service have the following root: apiRoot/apiWlanOffload/v1, where apiRoot and apiWlanOffload are discovered

AccessNetworkArea

3GPPAccessNetworkType

WLANAccessNetworkType

3GPPAccessNetwork

WLANAccessNetwork

ConnectivityDisable

ConnectivityEnable

ConnectivityAction

ConnectivityActionStatu

ConnectivityStatus

TracingAreaCode

PLMN

EUTRA_CellId

ConnectivityDisable

ConnectivityEnable

ConnectivityAction

ConnectivityActionStatu

ConnectivityStatus

ServiceSetIdentification

HemogeneousSSID

APIdentifier

38 “Е+Е”, vol. 53, 1-2, 2018

using service registry. The JSON content format is supported.

Table 1 provides an overview of the defined re-sources and applicable HTTP methods.

Table 1 Resources and аpplicable HTTP мethods

Resource name Resource URI HTTP method

AccessNet-workArea

/AccessNetworkArea GET

3GPPAccess NetworkType

/AccessNetworkArea/ 3GPPAccessNetworkType

GET

3GPPAccess Network

/AccessNetworkArea/ 3GPPAccessNetworkType/ 3GPPAccessNetwork

POST, GET, PUT, DELETE

WLANAccess NetworkType

/AccessNetworkArea/WLAN AccessNetworkType

GET

WLANAccess Network

/AccessNetworkArea/WLAN AccessNetworkType/ 3GPPAccessNetwork

POST, GET, PUT, DELETE

WLANAccess NetworkType

/AccessNetworkArea/WLAN AccessNetworkType

GET, PUT, DELETE

Connectivity Action

/AccessNetworkArea/ AccessNetworkType/ AccessNetwork/ ConnectivityAction

GET, PUT

Fig.3 shows a successful application-initiated of-

floading.

Fig.3 Successful application-initiated offloading to WLAN.

Fig.4 shows an unsuccessful application-initiated

offloading. Other possible responses indicating unsuc-cessful application-initiated offloading may be 401 Unauthorized (the Application did not submit creden-tials), and 404 Not Found (the Application provided an URI that cannot be mapped to a valid resource URI).

Fig.4 Unsuccessful application- initiated offloading to

WLAN.

Offloading state models

As a proof of the concept, we design models repre-senting the application, serving 3GPP access network and target WLAN. These models for a given UE have to be synchronized i.e. to expose equivalent behavior.

Fig.5 shows the mobile edge application view on the offloading state. In ConnectedTo3gpp state, the UE is connected to the serving 3GPP access network. Following a predefined policy, the Application may decide to initiate a UE offloading to the target WLAN. In ConnectedToWlan state, the UE successfully has performed the WLAN offload.

Fig.5 Application view on the offloading state.

Fig.6 shows the offloading state model as seen by the serving 3GPP access network. In 3gppConnected state, the UE is connected to the serving 3GPP access network. In HandoverToWlan state, the UE performs handover procedure to the target WLAN. In 3gppDisconnected state, the UE is disconnected from the 3GPP access network.

Fig.6 Offloading state model as seen by the serving 3GPP

access network.

Fig.7 shows the offloading state model as seen by the target WLAN. In WlanDisconnected state, the UE is not connected to the target WLAN. In Registering state, UE performs registration procedure, including authentication and session establishment. In Regis-tered state, the UE is registered to the target WLAN. In L3Configuration state, the UE acquires IP address.

App initiated offload Application

PUT …/AccessNetworkArea/WLANAccessNetworkType

/WLANAccessNetwork/ConnectivityAction (connectivityEnable)

200 OK

ConnectedTo3GPP

OffloadDecision/ OffloadToWlanReq

OffloadToWlanRes (success)

OffloadToWlanRes (failed)

OffloadingToWlan

ConnectedTo Wlan

3gppConnected

HandoverToWlan

OffloadToWlanReq /HandoverCommand

Release/ OffloadToWlanRes (success)

Connection re-establishment/ OffloadToWlanRes (failed)

3GppDisconnectApp initiated offload Application

PUT …/AccessNetworkArea/WLANAccessNetworkType

/WLANAccessNetwork/ConnectivityAction (connectivityEnable)

403 Forbidden

“Е+Е”, vol. 53, 1-2, 2018 39

In WlanConnected state, the UE is connected to the target WLAN.

Let us present the state machines as Labelled Tran-sition Systems (LTS).

Fig.7 Offloading state model as seen by the target WLAN.

Definition 1: A Labelled Transition System (LTS) is a quadruple (S, Аct, →, s0), where S is countable set of states, Act is a countable set of elementary actions, → ⊆ S × Act × S is a set of transitions, and s0 ∈ S is the set of initial states.

By ТApp = (SApp, АctApp, →App, s0App) it is denoted an

LTS representing the Application view on the offload-ing state, where: - SApp = ConnectedTo3GPP ( As1 ), OffloadingToWlan ( As2 ),ConnectedToWLAN ( As3 );

- ActApp = OffloadDecision ( At1 ), OffloadToW-lan(success) ( At2 ), OffloadToWlan(failed) ( At3 );

- →App = ( As1At2

As2 ), ( As2At3

As1 ), ( As2At2

As3 ); - s0

App = Connected. Short notations for states and actions are given in

brackets. By Т3GPP = (S3GPP, Аct3GPP, →3GPP, s0

3GPP) it is de-noted an LTS representing the serving 3GPP access network view on the offloading state, where: - S3GPP = 3gppConnected ( Os1 ), HandoverToWlan ( Os2 ), 3gppDisconnected ( Os3 );

- Act3GPP = OffloadToWlanReq ( Ot1 ), Release ( Ot2 ), ConnectionReestablishment ( Ot3 );

- →3GPP = ( Os1Ot1

Os2 ), ( Os2Ot3

Os1 ), ( Os2Ot2

Os3 ); - s0

3GPP = 3gppConnected.

By ТWLAN = (SWLAN, АctWLAN, →WLAN, s0WLAN) it is de-

noted an LTS representing the target WLAN view on the offloadfing state, where: - SWLAN = WlanDisconnected ( Ns1 ), Registering ( Ns2 ), Registered ( Ns3 ), L3Configuration ( Ns4 ), WlanCon-

nected ( Ns5 );

- ActWLAN = WlanRegisterationReq ( Nt1 ), WlanRegis-terationRes(success) ( Nt2 ), WlanRegisteration-Res(failed) ( Nt3 ), L3ConfigurationCommand ( Nt4 ),

L3ConfigurationComplete ( Nt5 ), L3ConfigurationFail

( Nt6 );

- →WLAN = ( Ns1Nt1

Ns2 ), ( Ns2Nt3

Ns1 ), ( Ns2Nt2

Ns3 ),

( Ns3Nt4

Ns4 ), ( Ns4Nt6

Ns1 ), ( Ns4Nt5

Ns5 ); - s0

WLAN = WlanDisconected. The mathematical concept of weak bisimulation is

used to prove that the three models expose equivalent behaviour.

Intuitively, in terms of observed behaviour, two LTSs are equivalent if one LTS displays a final result and the other LTS displays the same result. The idea of equivalence is formalized by the concept of bisimilarity [14]. In practice, strong bisimilarity puts strong condi-tions for equivalence which are not always necessary. In weak bisimilarity, internal transitions can be ignored.

Proposition: ТApp , Т3GPP and ТWLAN are weakly bi-similar.

Proof: To prove the bisimilarity between any two labelled transition systems, it has to be proven that there exists a bisimilar relation between their states. Let us denote by UAON a relation between sates of ТApp

, Т3GPP and ТWLAN where UAON=(ConnectedTo3GPP, 3gppConnected, WlanDisconnected), (OffloadingToWlan, HandoverToWlan, Registering, (ConnectedToWLAN, 3gppDisconnected, WLANConnected).

It is necessary to identify sequences of transition for the following states: from As1 , Os1 , Ns1 to As2 ,

Os2 , Ns2 ; from As2 , Os2 , Ns2 to As3 , Os3 , Ns5 ; and from As2 , Os2 , Ns2 to As1 , Os1 , Ns1 .

Then the following UAON relation exists for the events related to WLAN offloading: 1. In case of application-initiated offload to target

WLAN: for ( As1At2

As2 ) ∃ ( Os1Ot1

Os2 ) ⊓ ( Ns1

Nt1Ns2 ).

2. In case of successful application-initiated offload-ing to target WLAN: for ( As2

At2As3 ) ∃ ( Os1

Ot1Os1 )

WlanDisconnect

Registered

L3Configuration

L3ConfigurationCommand

WlanConnected

Registering

WlanRegistrationReq

WlanRegistrationRes (success)

WlanRegistrationRes (failed)

L3ConfigurationComplete

L3ConfigurationFail

40 “Е+Е”, vol. 53, 1-2, 2018

⊓ ( Os2Ot2

Os3 ), ( Ns3Nt4

Ns4 ), ( Ns4Nt5

Ns5 ). 3. In case of unsuccessful application-initiated of-

floading to target WLAN: for ( As2At3

As1 ) ∃

( Os2Ot3

Os1 ) ⊓ ( Ns2Nt3

Ns1 ) ⊔ ( Ns2Nt2

Ns3 ),

( Ns3Nt4

Ns4 ), ( Ns4Nt6

Ns1 ). Therefore ТApp, Т3GPPN and ТWLAN are weakly bisimilar.

Conclusion Multi-access Edge Computing appears to be an in-

tegrating technology for cellular and wireless technol-ogy as it may face the requirements for low latency, massive connectivity, and high reliability of a large number of applications.

In this paper, we present a new mobile edge service which provides open access to offloading functions in the radio access network. The service allows an au-thorized application to apply specific offloading policy based on terminal location, experienced QoS, requested data speeds, and the average number of call drops. With existing radio resource management mechanisms in different radio access networks it is the network or the device that initiates handover. The novelty of the re-search is the delegating the offloading control to third party applications. With the proposed functionality, an authorized application may first initiate a device pre-registration into the target access network and then based on measurements may initiate change of used radio access bearer.

Deployment of RESTful API for WLAN offloading enables applications aimed at efficient usage of limited radio resources and optimization of radio resource man-agement. For 3rd party application developers it is an opportunity to create new attractive applications, and for network operators it is a new source of revenue generation due to the reduction of data delivery costs.

Acknowledgements

The research is conducted under the grant of pro-ject DH07/10-2016, funded by National Science Fund, Ministry of Education and Science, Bulgaria.

REFERENCES

[1] Taleb, T., K. Samdanis, B. Mada, H. Flinck, S. Dutta and D. Sabella. On Multi-Access Edge Computing: A Survey of the Emerging 5G Network Edge Cloud Architecture and Orchestration. IEEE Communications Surveys & Tutorials, vol. 19, no. 3, pp. 1657-1681, third quarter 2017.

[2] Satyanarayanan, M. The Emergence of Edge Computing. IEEE Computer, January, 2017, pp.30-39.

[3] Rimal, B.P.D., P. Van, M. Maier. Mobile-Edge Computing vs. Centralized Cloud Computing over a Converged FiWi Access Network. IEEE Trans. on Network and Service Management, 2017, vol.PP, no.99, pp.1-1.

[4] Salmani, M., T. N. Davidson. Multiple access compu-tational offloading. 2016 IEEE 17th International Workshop on

Signal Processing Advances in Wireless Communications (SPAWC), Edinburgh, 2016, pp. 1-6.

[5] Flores, H., et al. Large-scale offloading in the Internet of Things. 2017 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), Kona, HI, 2017, pp. 479-484.

[6] Gehlot, A. Rajavat Handoff between WiMAX and WiFi wireless networks. Symposium on Colossal Data Analysis and Networking (CDAN), Indore, 2016, pp. 1-5.

[7] Velmurugan, T., Sibaram Khara, S. Nandakumar, and D. Sumathi. Seamless vertical handoff using modified weed optimization algorithm for heterogeneous wireless networks, radioelectronics and communications systems. 2017, Vol. 60, No. 10, pp. 431–448.

[8] Gohil, Z.M., D.N. Khandhar, K.M. Pattani, Implement & analysis of RSS threshold based vertical handoff decision algorithm for LTE & WLAN. International Journal of Innovative Research in Computer and Communication Engineering, vol. 5, issue 1, 2017, pp.311-317.

[9] Kumar, C., R.S. Shukla, G. Agarwal. Connectivity based handoff scheme for WiMAX. Int. Journal of Engineering Technology Science and Research IJETSR, vol.4, issue 7, 2017, pp.500-504.

[10] Sakthivel, S. A Spectrum access with user cooperation in heterogeneous cognitive radio networks. Int. J. of Systems Engineering, vol. 1, issue 3, 2017, pp. 63-66.

[11] Bayhan, S., L. Zheng, J. Chen, M. Di Francesco, J. Kangasharju, M. Chiang. Improving cellular capacity with white space offloading. 2017 15th International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt), Paris, 2017, pp. 1-8.

[12] 3GPP TS 23.402 Technical Specification Group Services and System Aspects, Architecture enhancements for non-3GPP accesses, Release 15, v15.1.0, 2017.

[13] Open Mobile Alliance. Diagnostics and Monitoring Trap Events Specifications, OMA-TS-DiagMonTrapEvents-V1_2-20131008-A, 2013.

[14] Pola, G., C. Manes, A. van der Schaft and M. D. Di Benedetto. Bisimulation Equivalence of Discrete-Time Sto-chastic Linear Control Systems. IEEE Transactions on Auto-matic Control, vol. PP, no. 99, pp. 1-1.

Prof. DSc. Evelina Pencheva is with Faculty of Tele-communications, Technical University of Sofia. She has a DSc degree in communication networks. Her scientific research area covers multimedia networks, telecommunication proto-cols, and service platforms. tel.:+359 2 965 3695 е-mail:[email protected].

Prof. DSc. Ivaylo Atanasov is with Faculty of Tele-communications, Technical University of Sofia. He received his DSc degree in communication networks and his scientific research area covers mobile networks, internet communica-tions and protocols, and mobile applications. tel.:+359 2 965 2050 е-mail:iiа@tu-sofia.bg

Assist. Prof. Eng. Denitsa Kireva is with Faculty of Tele-communications, Technical University of Sofia. Her scientific research covers IT and communications. tel.:+359 2 965 3544 е-mail:[email protected]


“Е+Е”, vol. 53, 1-2, 2018 41

INNOVATIVE TECHNOLOGIES

Innovation through Design Thinking, User Experience and Agile: Towards Cooperation Framework

Galia Nedeltcheva, Elena Shoikova

In the present paper are discussed the advantages and disadvantages of the methodologies of

Design Thinking, User Experience design and Agile аnd it is argued that their collaboration into practice is important as it leads to higher creativity, innovation and profitability. The authors are aware that there are quite a few books and articles written on how to perform Design Thinking, incorporate User Experience into the Agile process, innovate and apply them together into practice of projects. After a short introduction into the models, it is revealed how best to collaboratively implement Design Thinking principles and Agile practices by giving example with the IBM Design Thinking framework. The main advantages of the analyzed methodologies are that they have not been limited only to software development itself, they can be used by every type of business and can be implemented collaboratively at every possible level at the organization in order to generate more happy customers and manage to turn such happy customer into repeat customers.

Keywords – Agile Methodologies, Design Thinking, User Experience Design, Innovation

Иновации чрез синергията на проектантско мислене, потребителско изживяване и гъвкави методологии (Галя Неделчева, Елена Шойкова). Настоящата статия разглежда предимствата и недостатъците на методологиите Проектантско/Дизайнерско мислене (Design Thinking), Проектиране/Дизайн на потребителския опит/изживяване (User Experience design) и Гъвкавите методологии (Agile methodologies). В нея се дискутира взаимоотношени-ето, при което полученият ефект е по-голям от сумата на индивидуалните ефекти на тези методологии в практиката, като важна предпоставка за тяхната по-голяма креативност, иновации и ползи. Авторите са наясно, че има издадени малко книги и статии, които дават яснота как да се приложи методологията на дизайнерското мислене и потребителския опит в гъвкавия процес (Agile), как да се направи иновация и как те могат да се приложат на прак-тика в проектите. Структурата на изложението е следната: кратко въведение в тези моде-ли; предложения как най-добре да се приложат съвместно принципите на дизайнерското мис-лене и практиките на гъвкавите методологии; практически пример с рамката на дизайнерско мислене, предложен от IBM. Основните предимства на анализираните методологии са, че те не са ограничени само до приложение в софтуерната разработка, а могат да бъдат използва-ни на практика от всеки тип бизнес и на всяко организационно равнище, за да привлекат пове-че задоволени клиенти и да ги накарат да останат за дълго.

Introduction Design Thinking (DT) is a people-centered model

that encourages creativity and innovation to create a product or service that solves a complex problem for the target customer or user. It can help innovate a new product, design a simple solution to a complex problem, or to get the whole team involved in generating design ideas [1].

DT is a more strategic, problem-solving, big-idea process instead, while User Experience (UX) is based

on a more tactical work on a digital experience. UX is grounded also in doing empathic research to help define the problems to solve, find ways to innovate, and is used to continue designing, testing, iterating, and building a product or service. The two methodologies are almost the same, although the collaboration aspect of DT is not always done in the UX process in practice. However, collaboration is a very important ingredient on an Agile software project and an effective way to conduct design. Therefore, DT fits well into the Agile model of doing digital design.

42 “Е+Е”, vol. 53, 1-2, 2018

There are quite a few books and articles written on how to perform DT and its phases (see Figure 1), how to incorporate UX into the Agile process, how to innovate and how to apply them into practice in a project [1].

Fig. 1. Phases in the Design Thinking process [1].

Agile software development leads to improved software delivery process by using short and fast iterations. By simplifying the software engineering practices Agile methodology could guarantee a higher success rate than traditional waterfall model [10]. The Agile Manifesto requires collaboration with customers, but its approach does not guarantee that the software development team will work towards solving the correct problem [11]. Agile methods including Scrum have already tried to follow this approach by incorporating user feedback as part of the requirement process.

In the Agile software development it is important to understand the end user needs, to incorporate the user feedback into the software development process. The user stories should represent the need of the user not only the view of the software development team. Moreover, IBM DT extends the original DT method providing a new approach to write requirements, organize teams and track project progress including end-user feedback during all the project development phases [12].

The contributions of this paper are as follows: (i) to point out the advantages and disadvantages of the DT, UX design and Agile single methodology and the need for their collaborative integration. (ii) To present shortly the best practices of IBM DT and Autodesk use case as well as to draw a conclusion. After the short introduction, the structure of the paper is as follows: Section 2 gives a short background on DT and Agile methodologies. Section 3 discusses the possible coupling of DT and Agile, presenting the strength of Agile and the single phases of the DT. Section 4 outlines the best practices and ideas for employing DT on an Agile project, in particular the IBM DT framework and Autodesk use case. Section 5 draws conclusions about how best to implement DT principles and Agile practices as well as what exactly makes UX-focused companies successful.

Background on Design Thinking and Agile methodology

Design Thinking is a methodology for innovation that combines creative and analytical approaches and requires collaboration across disciplines.

The focus is on creating spectacularly transformative learning experiences and along the way, students develop a process for producing creative solutions to even the most complex challenges they tackle [14].

Empathise, define, ideate, prototype, test There are five key phases and principles to

successful DT [2], as follows: • Empathize with your audience, customers or

target market • Define their needs, any problem they might

have and your understandings • By challenging assumption one is able to

ideate or create ideas for innovative solutions • Look at a prototype and start creating

solutions • Finally, it is imperative to test these solutions

The principles of DT appear intuitive. What

distinguishes this approach from human-centric innovation is how each of the steps has been conceptualised [14].

Fig. 2. The principles of Design Thinking [14].

Empathise: Design Thinking is grounded on a deep understanding of the people you are trying to serve. It requires careful observation of people within their contexts to uncover disconnects between what people say and what they do which is where great insights can often be found. Design thinkers also engage with people in deep and meaningful ways through loosely structured conversations. And of course, they listen and watch.

“Е+Е”, vol. 53, 1-2, 2018 43

In many cases, the best solutions are the ones that address the needs of the ‘extreme user.’ During a recent visit to the d.school by participants of the 2016 FIPP/VDZ Innovators’ Tour, Astrid Maier, a journalist and Knight Fellow, described the example of carry-on luggage which was initially designed to meet the unique needs of airline pilots but which today is used by virtually every traveler.

Define: After developing a deep understanding of the user, the next step is to define the challenge you are taking on based on what you have learned about your user and about the context. This involves creating a meaningful and actionable problem statement that guides the rest of the process. The problem statement should address who the user is, what specific needs they have and what insights the team had during the empathise stage.

Ideate: This is the phase where the magic happens in Design Thinking because the team comes together to generate the broadest range of ideas they can in a judgement-free environment. Successful ideators build on each other’s ideas while setting some parameters to ground the discussion. Situating an ideation session in a creative space that includes inspiring materials can help the process.

At the d.school, multi-colored post-it notes are often a key enabler of this process as they help the team visualize an idea. Other brainstorming techniques include bodystorming, which helps designers discover new and unexpected ideas by physically experiencing a situation; mindmapping; and sketching.

Prototype: Once the team has identified one or more solutions to the problem statement, it is time to develop prototypes. This is an iterative exercise where ‘artifacts’ are created and tested in attempt to get closer to a final solution. The key to prototyping is creating something fast and cheaply that users can interact with. It is also critical to identify what’s being tested with each prototype and what design question is being addressed.

The prototyping area at the d.shool is stocked with very simple materials – stacks of Post-it note pads, paper clips, staplers, cardstock and only a few power tools.

Test: The only reason to create prototypes is to have the opportunity to get feedback that allows you to understand your user better. Testing will yield the best insights if users are able to experience the prototype in a real way with very little upfront explanation. The job for the designer is to observe how the prototype is used or misused, how the user interacts with it and what they say about it.

Outside of the d.school, media organisations have a number of tools at their disposal for rapid prototyping and testing – including A/B testing of content and ad formats, putting video on Youtube, social media posts and ads and many more. This allows for testing a wide range of new ideas fast and eliminating those that don’t deliver results.

Iterate

The unspoken but implicit next step or intra-step in DT is iteration. This may involve repeating the entire process a number of times before arriving at a solution to the problem statement that the team loves as well as creating and testing multiple prototypes.

The principles of DT provide a rich opportunity for media companies to embed innovation deeply into their cultures. By focusing more on creating innova-tors rather than innovations, Design Thinking encour-ages organisations to place their bets on the collective creativity of their people as they navigate the twin challenges of uncertainty and opportunity [14].

These phases are not necessarily sequential instead, they should be used organically without any worry about a specific order in which they should occur, with the solution being the end-goal rather than the way it was achieved.

DT is the most effective way to understand the customer’s needs and it involves finding out what and how they do and then thinking out of the box or coming up with crazy ideas that might just be the solution to a unique offering that customers will love. While DT provides the solution, the next aspect is how to deliver those solutions to the marketplace.

Agile works by focusing on smaller bits of the larger project at a time, delivering value quickly and adapting or changing based on real-time feedback. Scrum is a derivative of the Agile that follows the same iterative process and is used by Fortune 500 companies world wild. It is an effective, smart and perfect example of how maximum input equals maximum output. The following is a quick overview of how Scrum works: • Sprint: this is simply a team planning meeting to

determine what needs to be done in the coming sprint.

• Daily stand-up: or a daily scrum, is a 15 minute get together or mini-meeting for everyone to touch base.

• Sprint demo: this is a sharing meeting where the team shows what they’ve achieved in that sprint.

• Sprint retrospective: this is a review of what worked and what didn’t, what went well and what needs improvement and actions for the next sprint to be better.

44 “Е+Е”, vol. 53, 1-2, 2018

While DT helps us understand what work to do, Scrum gives us the autonomy to decide how to do it. The main similarity is that both DT and Scrum are iterative and the mentality and approach required to successfully undertake both are quite similar. They both require adopters to develop sufficient insight in order to recognize early successes and failures through constant evaluation and adaptation.

The DT team need not view their work as completely independent of the project team or vice versa. DT encourages the cross-functional team to work together and that team could include developers, user experience designers and testers [2].

Like DT, Agile emphasizes collaboration, but it has its own language and methodologies. The Agile manifesto outlines some key principles and then teams typically use methodologies like Scrum, Sprints, Retrospectives to execute.

Agile's iterative approach lets project teams react to changes and challenges better and deliver a finished product in a shorter time frame. However, developing the product faster is good only if it is the right destination to begin with. Organiations are coupling DT and Agile processes aiming at identifying the right solution and then focusing on building a better product. Understanding the problem paves the way for executing more creative and useful solutions through Agile processes, i.e. delivering the true customer satisfaction on which Agile is built [3].

Integrating DT and Agile approaches helps organizations find and build the right customer-focused solution.

Agile uses the DT approach in identification of problems and challenges, and in generation of ideas and innovative solutions. The process includes series of steps and tools that proposes an analysis and discussion of the current situation from a holistic point of view. Once the real problem is identified, the focus turns to the solution, guided by the end user’s needs by stimulating “outside of the box” perspectives. All this is possible through participation of teams who collaborate within an environment that encourages creativity. Agile is qualified to help a company using DT to obtain solutions that apply new models and meet business challenges.

This methodology searches different ways for the innovative solution to the problem, prioritizing collaborative work in multidisciplinary teams [6]. Summing up the benefits of Agile are as follows: • Faster and cheaper process to generate innovation; • Focus on customer perception, their needs, desires

and behavior;

• Process based on implicit knowledge and experiences with prototypes;

• Innovation as a result of the use of methodologies plus teamwork;

• "Outside the box" perspectives

Collaboration between Design Thinking, User Experience and Agile

DT is a strategy for product design that avoids the mistakes of extrapolating future customer requirements based on past data or defining a strategy based on instinct instead of the actual needs. So that, in DT we shift the focus to human behavior.

While DT provides a robust, creative and innovative process it does not necessarily address actual product development limitations and requirements from an organizational perspective. Thus, theoretically we can iterate through the 5 steps with little involvement from the engineering, product development and operational units. We might segregate and silo the design – 5 step process and end with a lengthy requirement document that feeds to a linear approach such as waterfall for the actual development. Otherwise, we can view the DT process as a sequential process in itself and superimpose it on the existing development methodology, encumbering the creative innovative process with a procedural methodology. We end up with product iterations that are 9 months long which are detrimental to fast feedback loops.

Agile development is a term which refers to several iterative and incremental development processes that speed up the delivery of the business value. Through a process of continuous planning and feedback it ensures that value is maximized throughout the development process. The most popular include Extreme Programming (XP), Scrum, Scaled Agile, Dynamic Systems Development Method (DSDM), Lean Development, and Feature-Driven Development (FDD). Agile methods are those that follow the Agile Manifesto and Agile principles [15]. While Agile provides a paradigm shift in how products are delivered, it does not necessarily address the challenge of what to deliver. The product backlog, which is a common artefact in many Agile processes, is populated by an invisible process outside the product organization. Practically many Agile transformations do not address new product and strategy initiation at all. It is not uncommon that users use less than 20-30% of the product features. This results in 70-80% of the development effort being a waste.

“Е+Е”, vol. 53, 1-2, 2018 45

Agile and DT are complementary. DT is particularly well-suited to situations where the problem itself is not clear, focusing strongly on problem definition, problem shaping, and requirements clarification. Likewise, Agile methods embrace uncertainty and are appropriate for projects where the requirements are subject to change [7].

While DT is a solution-centric approach, it also places great emphasis on having a clear understanding of the problem. For Agile projects, the backlog is where the functional requirements of the system under development are captured, and the quality of those requirements is a significant factor determining the success of the project [7].

A pairing of DT with an Agile mindset and method can occur across the lifecycle of a project, from Initiate to Release phase (see Figure 3). Examples of how to elevate the Agile with DT [7] are as follows:

• Ideation to develop the Agile backlog • Empathy puts the ‘human’ at the centre of the

Agile backlog • Creativity and innovation reinvigorate Agile

teams

Fig. 3. Agile with Design Thinking: touchpoints.

DT is an exploratory approach which relies on identifying end-user needs and discovering solutions to meet those needs. UX designers are generally better appointed when it comes to think from the end-user’s point of view. Their job is to help in the development of solutions which are easy to use. That is why being Agile goes beyond the use of tools and the designers need to understand their importance in the team and should be actively involved in providing design suggestions for the improvement of the software at all stages.

Sometimes a small change in the interaction design can reduce a lot of problems faced by the developers.

However such changes in mobile app design for instance are usually accompanied by an increase in the development work. Agile will need such changes being an iterative process where all those changes that help in achieving a business goal are always welcome. That is why Agile is so effective in delivering quick value so that, it allows you to adapt quickly based on feedback and learning. This also means that the work of a UX designer is never over in an Agile project.

Best practices and ideas for employing design thinking on an agile project

IBM Design Thinking framework

IBM DT approach is a framework for teaming and action and it helps the teams to deliver outcomes. So that, it is a set of recommended practices that helps one moves beyond feature-centric delivery. In order to implement the Software Development Framework (SDF) successfully it is necessary to create multidisciplinary teams composed by designers, engineers, product managers and users who work together to drive a vision of the software development [9].

Some limitations of the IBM DT process are related to the project team structure. The process itself could not be applied successfully if the company does not change its approach to solving problems. Implementing IBM DT requires teams to reorganize as well as review their work model and functional roles.

Fig. 4. IBM Design Thinking compared to traditional

Design Thinking [9].

IBM Design Thinking SDF goal is to extend DT principles so they can be applied to develop software that captures user needs at the speed and scale required for fast pace incremental software development such as Cloud based software. While it shares some similarities with other DT methods, it has a few modifications, including three practices that are unique to the framework: sponsor users, playbacks and hills [12]. The integration of those practices to DT are shown in Figure 4.

46 “Е+Е”, vol. 53, 1-2, 2018

The IBM DT defines three major roles with different set of responsibilities. • The Product Manager is responsible for

understanding the market opportunity and defining the product release. He is responsible for the project kick-off, defining and recruiting sponsor users, and defining the playback strategy.

• The Designer is responsible for the UX and functional design. He is engaged in developing design artifacts, mock ups, user research and design sprint plan.

• The engineering team is responsible for the technical design and implementation of the release. They are in control of project architecture and executable code, prototype and the technical sprint plan.

IBM Design Thinking SDF activities are divided into two main phases. The Visioning phase is responsible to develop software requirements through the use of several DT practices that combines user personas, empathy maps, hills and story maps. The other phase, the Delivery Wave consists of software development Sprints conducted by multidisciplinary teams that includes Sponsors Users, who contribute with constant feedback about the delivered artifacts. Figure 5 illustrates the workflow of a sample IBM DT project [9].

Summing up, the IBM DT principles are as follows [13]:

• Diverse empowered teams • A focus on user outcomes • Restless reinvention • See problems and solutions from a new point

of view

Fig. 5. IBM Design Thinking software development

framework [9].

The Loop consists in: Understand users’ needs and deliver outcomes

continuously

At the heart of IBM DT is a behavioral model for understanding users’ needs and envisioning a better future: a continuous loop of observing, reflecting and making [13].

IBM DT incrementally delivers great UX, while Agile incrementally delivers great enabling software. What links them most closely is the continuous cycle of experience maps and playbacks.

DT facilitators initiate and lead DT activities on their team to reach great outcomes for their users. With time and practice anyone can become an effec-tive and credible facilitator. IBM DT is designed as a framework and the team could use bits and pieces of it as it makes sense. As a DT facilitator one should en-sure that conversations and activities are centered on the user.

Each activity helps to understand the user’s prob-lems and motivations, explore new concepts, proto-type designs and evaluate with stakeholders. These activities serve as tools for establishing the IBM DT framework [13], which are as follows:

Stakeholder Map

If you’re integrating new team members, starting a new project, exploring a new market, or expanding an offering, this activity helps you identify project stake-holders, their expectations, and relationships.

Empathy Map - It helps to rapidly put the team in the user’s shoes and aligns whether at the beginning of a project or mid-stream when one needs to re-focus on the user.

Scenario Map (As-is/ To-be) - As-is Scenario Maps help documented collective understanding of user workflows and are best used as precursors to explor-ing new ideas. To-be Scenario Maps tell the story of a better experience for the user.

Big Idea Vignettes - Once the team has a clear and validated understanding of the user’s problems and challenges, this activity is a good way for many peo-ple to rapidly brainstorm possible ideas.

Prioritization Grid - When many items (such as ideas, Hills, scenarios, or user stories) are being con-sidered, this activity helps the team to evaluate and prioritize them by focusing discussions on importance and feasibility.

Needs Statements

This is a very effective activity to use with the team when one feels that he is shifting from the actual needs, desires and goals of the user. It helps focusing the work around the user.

Storyboarding - is a way to iterate and communi-cate ideas and scenarios visually by telling user-centric stories.

“Е+Е”, vol. 53, 1-2, 2018 47

Assumptions and Questions - Any time you feel that your team’s work needs a “reality check,” use this activity to identify and prioritize what assumptions are being made, what you’ve been guessing about, and what your team still doesn’t know.

Feedback Grid - This activity helps to gather and organize any sort of feedback and to unpack questions and ideas either in real time or after-the-fact as an efficient means of determining next steps.

Experience-based Roadmap - This activity helps you define a “minimum delightful experience” by scoping big, visionary ideas into more achievable near-term outcomes while still focusing on the user experience.

Discovery-driven enterprise user-experience de-sign: case of Autodesk

At Autodesk the company’s design teams are as global as its customer base. The $2.5 billion-dollar software giant is powered by 7700 employees across all 7 continents [14].

In the Tel Aviv office, Uri Ashano serves as the senior UX manager for AutoCAD 360, the mobile application of the company’s flagship product. The AutoCAD® mobile app is a drawing and drafting tool that lets you view, create, edit, and share AutoCAD drawings on mobile devices—anytime, anywhere.

Uri Ashano and his team of five (two UX design-ers, two visual designers and one researcher) collabo-rate closely with the San Francisco HQ to practice user-centered design within an Agile process.

As Uri explains, the company sees itself as a knowledge house, not just a software provider. All designers train and work with the Luma Innovation Institute, which teaches 36 different methods for user-centered design. The design process Uri and team follow each time they get new feature requests, illu-minates the power of collaborative design, especially in the discovery stage.

The Tel Aviv team’s process begins when they get a request from the larger AutoCAD 360 team for a new feature. In response, the team first opens a new project in Slack to start investigating the problem the feature aims to solve. This initial research consists of interviewing local architect firms, reviewing customer support tickets for ideas, and reviewing online data in MixPanel. The team also consults data from the great-er Autodesk’s research, mostly around multiplatform use of AutoCAD and the flows related to these products.

The request usually presents itself as a user scenar-io such as: “An architect needs AutoCAD drawing at his job site. He’s bringing along an iPad (or other

tablet) and wants to view his drawings, modify and annotate. Afterwards, he wants to share the updates with colleagues.”

In summary, Autodesk’s Tel Aviv team shows that enterprise design does not need to be bogged down by poor communication and mountains of documenta-tion. Here are the big takeaways:

• Validate feature requests with early qualita-tive research (reviewing support tickets and user interviews) and quantitative research in analytics tools and surveys.

• Dedicate time up front for discovery and idea-tion with half- to full-day workshops aimed at defining the rough feature set.

• Treat documentation as a knowledge portal rather than a paper trail by linking out to fur-ther details.

• Prioritize prototyping for the most difficult in-teraction models [14].

Conclusion

The integration of DT, UX and Agile offers great advantages for innovation, creativity and profitability, and they are not limited only to software development itself, but they can be used by every type of business. More importantly, the benefits associated with each one can be significantly increased by combining them to create a process that works seamlessly.

By combining both DT, UX and Agile, the team could recognize a couple of advantages, as follows: • The Scrum development teams form part of the

design process from the beginning, which enables them to understand in-depth the customer’s needs and requirements. This way a connection is built between the two processes, which can improve the product.

• Combining both processes help to eliminate duplication and create products or services which meet the actual user needs.

• The iterative and user oriented software development process connects with the creative, iterative and user oriented DT and UX processes and this way is created a collaborative environment at the organization.

• Through the application of a combination of DT, UX and Agile approaches, organizations can reduce risks from the development effort and ob-tain better return on investment.

Тhe results of our study show that UX practitioners involved in highly effective projects apply research and design-thinking techniques to inform product design, and lead activities that foster team cohesion and collaboration. They conduct both generative and

48 “Е+Е”, vol. 53, 1-2, 2018

evaluative research to identify feature gaps and examine both long-term and short-term solutions. Rather than just waiting for feature requests to come from product managers and stakeholders, UX people proactively reach out to end users, discovering opportunities for their products to fill unmet user needs and stand out in crowded markets.

Although, UX development and user centered de-sign practices have been followed by many organiza-tions (Apple, Coca-Cola, Ford, IBM, Intuit, Procter & Gamble, Starbucks, Walt Disney and Nike [16]) to enhance their product efficiency and user satisfaction, they have always been considered a formality, and have never been able to fully incorporate them into Agile development processes, due to a string of chal-lenges discussed in this study.

There is a need to conduct a much larger scale re-search study such as an extensive literary review and a survey study across many organizations that have varying development methodologies, of different scales and competencies, in order to get a better un-derstanding of the challenges that plague the integra-tion process. This will help generate more concrete frameworks and generalized theories, in order to help the industry to adopt the best practices.

Acknowledgments

The research is supported by the KoMEIN Project (Conceptual Modeling and Simulation of Internet of Things Ecosystems) funded by the Bulgarian National Science Foundation, Competition for financial support of fundamental research (2016) under the thematic priority: Mathematical Sciences and Informatics, con-tract DN02/1/13.12.2016.

This article was reported at the TELECOM 2017 Conference.

REFERENCES

[1] 6 Best Practices for Integrating Design Thinking into an Agile Project https://www.tandemseven.com/ blog/6-best-practices-integrating-design-thinking-into-agile-project/

[2] Design Thinking and Agile – A Great Combina-tion for Effective Innovation http://www.corporate-alchemists.com/design-thinking-agile-innovation/

[3] Burba, D. Agile by Design: Integrating Design Thinking and Agile Approaches Helps Organizations Find and Build the Right Customer-Focused Solution. PM Net-work, 30(10), 2016, pp. 58–63.

[4] Design Thinking and Agile – Complimentary and Powerful, https://sapir-cs.com/uncategorized/design-thinking-and-agile-complimentary-and-powerful/ Design Thinking in Agile: Google Design Sprint vs. IBM Design Loop, https://socaluxcamp.com/sessions-2017/?design-thinking-in-agile-google-design-sprint-vs-ibm-design-loop

[5] Design Thinking - A New Look to Transform In-novation into Reality, http://agilesolutions.com/release/design-thinking/

[6] Elevate your Agile with Design Thinking, http://blog.deloitte.com.au/elevate-your-agile-with-design-thinking/

[7] The Role of Design Thinking in Agile Environ-ments, http://www.mindhatchllc.com/insights/ design-thinking-agile/

[8] Percival Lucena, Alan Braz, Adilson Chicoria, and Leonardo Tizzei. IBM Design Thinking Software Devel-opment, Framework. Springer International Publishing AG 2017 T. Silva da Silva et al. (Eds.). WBMA 2016, CCIS 680, pp. 98–109, 2017. DOI: 10.1007/978-3-319-55907-0 9

[9] VersionOne. Annual State of Agile Development Survey. VersionOne Inc., 2016.

[10] Norman, D.A. The Design of Everyday Things: Revised and Expanded Edition. Basic books, New York, 2013.

[11] Azis, M. IBM Design Thinking (2016a). http://www.ibm.com/design/thinking/

[12] IBM Design Thinking Field Guide. IBM Corpora-tion, 2017.

[13] Jerry Cao. How Autodesk Tackles UX Design Challenges, https://www.uxpin.com/studio/blog/how-autodesk-tackles-ux-design-challenges/?utm_source=email-campaign&utm_medium=email&utm_campaign=trial-nurture&utm_content=agile&mkt_tok=eyJpIjoiWVRBMU5HUmpNRGRoTWpjeSIsInQiOiJMZjRUVk1CcU9JZFRJQXlqd0dVODEyUkdvU010R21UUjN

[14] Ambysoft Inc. Examining the Agile Manifesto. 6 April 2011. Design Management Institute, Feb 2017 http://www.dmi.org/blogpost/1093220/182956/Design-Driven-Companies-Outperform-S-P-by-228-Over-Ten-Years--The-DMI-Design-Value-Index

Elena Shoikova, DScTech. Eng., Professor with Com-puter Science Department at the University of Library Stud-ies and Information Technologies, Bulgaria. Her scientific interests are mainly in the area of Software Engineering, Internet of Things, Design Thinking, User Experience De-sign, Virtual Reality, Augmented Reality, Agile, E-learning Methodology & Systems and Smart Learning.


Galia Nedeltcheva, PhD, Assist. Professor with the Faculty of Mathematics and Informatics at Sofia Universi-ty, Bulgaria. Her scientific interests are mainly in the field of Software Engineering, Cloud Computing, Internet of Things & Big Data, Design Thinking, User Experience Design and Agile.

e-mail: [email protected].


“Е+Е”, vol. 53, 1-2, 2018 49

ЕЛЕКТРОТЕХНИКА И ЕЛЕКТРОНИКА E+E 53 год. 1-2/2018 Научно-техническо списание Издание на: Съюза по електроника, електротехника и съобщения /CEEC/

Главен редактор: Проф. дтн Иван Ячев, България

Зам. гл. редактор: Проф. дтн Сеферин Мирчев, България Редакционна колегия: Проф. д-р Анатолий Александров, България Проф. д-р Венцислав Вълчев, България Д-р Владимир Шелягин, Украйна Проф. д-р Георги Стоянов, България Доц. д-р Евдокия Сотирова, България Доц. д-р Захари Зарков, България Проф. Кристиан Магеле, Австрия Проф. Маурицио Репето, Италия Проф. д-р Марин Христов, България Проф. дтн Михаил Анчев, България Проф. д-р Никола Михайлов, България Проф. дтн Ради Романски, България Проф. д-р Росен Василев, България Проф. Такеши Танака, Япония Проф. Ханес Топфер, Германия Д-р Хартмут Брауер, Германия Акад. проф. дтн Чавдар Руменин, България Акад. проф. Юрий Якименко, Украйна Проф. Юън Ричи, Дания Консултативен съвет: Чл. кор. проф. дфн Георги Младенов, България Проф. д-р Димитър Рачев, България Проф. дтн Емил Соколов, България Проф. д-р Жечо Костов, България Доц. д-р Иван Василев, България Проф. дтн Иван Доцински, България Доц. Иван Шишков, България Проф. дтн Людмил Даковски, България Проф. дтн Минчо Минчев, България Проф. дфн Николай Велчев, България Доц. д-р Петър Попов, България Проф. дтн Румяна Станчева, България Проф. д-р Стефан Табаков, България Технически редактор: Захари Зарков Адрес: ул. “Раковски” 108 ет. 5, стая 506 София 1000 тел.: +359 2 987 97 67 e-mail: [email protected] http://epluse.fnts.bg

ISSN 0861-4717

С Ъ Д Ъ Р Ж А Н И Е

ЕЛЕКТРОНИКА Марин Б. Маринов Мобилна система за мониторинг на фини прахови частици 1

Кирил Иванов, Иво Илиев, Серафим Табаков Плазмена стерилизация, чрез нов многоточков аксесоар 10

Валери Петков, Николай Маджаров Проектиране на аксиален ротационен безконтактен предавател на енергия 16

Виктор Томов, Иво Илиев, Весела Кръстева Анализ на формата на изходния ток за контрол върху процеса на искрене в електрохирургически апарати 22

Галя Маринова, Здравка Чобанова Изследване на факторите влияещи на консумацията на проекти, реализирани върху програмируеми схеми FPGA 28

ТЕЛЕКОМУНИКАЦИИ Евелина Пенчева, Деница Кирева, Ивайло Атанасов Услуга за множествен достъп за приложения, иницииращи разтоварване 35

ИНОВАТИВНИ ТЕХНОЛОГИИ Галя Неделчева, Елена Шойкова Иновации чрез синергията на проектантско мислене, потребителско изживяване и гъвкави методологии 42 През 2018 г. списание Electrotechnica & Electronica се издава с финансовата подкрепа на фонд „Научни изследвания“ към Министерството на образованието и науката.

INT

’L P

RO

GR

AM

CO

MM

ITT

EE

Pro

f. D

ench

o B

AT

AN

OV

, P

h.D

.

(Tha

iland

) P

rof.

Fra

nces

co B

ER

GA

DA

NO

(Ita

ly)

Pro

f. D

umitr

u D

an B

UR

DE

SC

U, P

h.D

.

(R

oman

ia)

Ass

oc. P

rof.

Vas

sil F

OU

RN

AD

JIE

V, P

h.D

.

(Bul

garia

) P

rof.

Van

gel F

US

TIK

, D.S

c.

(Mac

edon

ia)

Ass

oc. P

rof.

V.

GA

YO

SO

MA

RT

ÍNE

Z, P

h.D

. (S

pain

) P

rof.

Iliya

GE

OR

GIE

V, P

h.D

.

(U

SA

) A

ssoc

. Pro

f. A

lain

e M

. G

UIM

AR

ÃE

S, P

h.D

. (B

razi

l) P

rof.

Luis

HE

RN

ÁN

DE

Z E

NC

INA

S,

D.S

c.

(S

pain

) P

rof.

Ata

nas

ILIE

V, P

h.D

.

(M

aced

onia

) A

ssoc

. Pro

f. Iv

an J

ELI

NE

K, P

h.D

.

(Cze

ch R

epub

lic)

Dr.

Kar

l O. J

ON

ES

(U

K)

Dr.

Zea

shan

H.

KH

AN

(Pak

ista

n)

Pro

f. D

ong

Hw

a K

IM, P

h.D

.

(K

orea

) P

rof.

Kar

ol M

AT

IAŠ

KO

, Ph.

D.

(S

lova

kia)

A

ssoc

. Pro

f. B

rojo

MIS

HR

A, P

h.D

.

(I

ndia

) A

ssoc

. Pro

f Jez

abel

MO

LIN

A-G

IL

(Spa

in)

Ass

oc. P

rof.

Irin

a N

ON

INS

KA

, Ph.

D.

(Bul

garia

) P

rant

osh

Kum

ar P

AU

L, P

h.D

.

(Ind

ia)

Pro

f. E

velin

a P

EN

CH

EV

A, D

.Sc.

(Bul

garia

)

Pro

f. G

oran

RA

FA

JLO

VS

KI,

Ph.

D.

(Mac

edon

ia)

Pro

f. R

adi R

OM

AN

SK

Y, D

.Sc.

(C

hairm

an)

(B

ulga

ria)

Pro

f. Le

on R

OT

HK

RA

NT

Z,

D.S

c.

(T

he N

ethe

rland

s)

Ass

oc. P

rof.

Sud

hir

K. R

OU

TR

AY

, Ph.

D.

(I

ndia

) P

rof.

Hea

ther

RU

SK

IN, P

h.D

.

(I

rela

nd)

Pro

f. E

ugen

e S

EM

EN

KIN

, D

.Sc.

(Rus

sia)

P

rof.

Rad

omir

ST

AN

KO

VIĆ

, Ph.

D.

(Ser

bia)

P

rof.

Ivan

TA

SH

EV

, Ph.

D.

(US

A)

Ass

oc. P

rof.

Dim

itar

TS

AN

EV

, Ph.

D.

(Bul

garia

) P

rof.

Kal

um P

. U

DA

GE

PO

LA, P

h.D

.

(A

ustr

alia

) P

rof.

Laris

sa Z

AIT

SE

VA

, D

.Sc.

(Lat

via)

O

RG

AN

IZA

TIO

N

Th

e I

nte

rnati

on

al

Co

nfe

ren

ce o

n I

nfo

rmati

on

Tech

no

l-o

gie

s (

Info

Tech

) is

a s

uccessor

of

the I

nte

rnational C

on-

fere

nce o

n S

yste

ms for

Auto

mation

of E

ngin

eering a

nd R

e-

searc

h (

SA

ER

), e

sta

blis

hed in 1

987.

Info

Tech

-2018 w

ill be org

aniz

ed as a c

om

bin

ed f

orm

at:

[1]

Pre

limin

ary

dis

cussio

n o

n s

ubm

itte

d r

eport

s b

y u

sin

g

virtu

al e-f

oru

m o

n the w

eb s

ite;

[2]

Pre

senta

tion o

n s

ite (

confe

rence s

essio

ns)

of all

report

s

accepte

d t

o b

e p

ublis

hed in C

onfe

rence P

roceedin

gs.

Part

icip

ants

should

send a

ll m

ate

rials

to the e

-mail

addre

ss:

rrom

@tu

-sofia.b

g

MA

IN D

EA

DLIN

ES

till

30.0

4.20

18

Sen

ding

the

App

licat

ion

For

m a

nd f

ull p

aper

in E

nglis

h.

Sta

rtin

g th

e R

evie

win

g pr

oces

s.

till

30.0

5.20

18

Rep

orts

rev

iew

ing

and

acce

ptan

ce c

onfir

mat

ion

by s

endi

ng

Sec

ond

For

m a

nd R

evie

wer

’s R

epor

ts.

till

30.0

6.20

18

Reg

istr

atio

n fo

r pa

rtic

ipat

ion

by F

ee p

aym

ent,

fill i

n th

e R

egis

tra-

tion

For

m a

nd R

evis

ed p

aper

sub

mis

sion

.

15.0

7 –

15.0

8.20

18

Virt

ual e

-For

um o

rgan

izat

ion

with

dis

tanc

e di

scus

sion

.

20 –

21.

09.2

018

Con

fere

nce

open

ing

and

repo

rts

pres

enta

tion

in s

essi

ons.

PA

RT

ICIP

AT

ION

FE

ES *

Par

ticip

atio

n w

ith 1

rep

ort

120

BG

N

or

65 E

UR

For

eac

h ad

ditio

nal r

epor

t

50 B

GN

or

25

EU

R

Par

ticip

atio

n w

ithou

t rep

ort

and

Acc

ompa

nyin

g pe

rson

70 B

GN

or

35

EU

R

* P

aym

ent

bef

ore

30

Jun

e

Th

e P

arti

cip

atio

n F

ee in

clud

es C

onfe

renc

e m

ater

ials

(P

rogr

am,

Pro

ceed

ings

, et

c.),

par

ticip

atio

n in

all

conf

eren

ce s

essi

ons,

cof

-fe

e br

eaks

, C

onfe

renc

e D

inne

r, d

isse

min

atio

n an

d se

ndin

g to

E

BS

CO

dat

abas

es.

20%

VA

T i

s in

clu

ded

.

OR

GA

NIZ

ING

CO

MM

ITT

EE

Cha

irman

: R

adi R

OM

AN

SK

Y

Mem

bers

: D

imita

r T

ZA

NE

V, I

rina

NO

NIN

SK

A,

Tod

or

KO

BU

RO

V,

Ele

na P

AR

VA

NO

VA

, Des

isla

va IV

AN

OV

A

INF

OR

MA

TIO

N

Pro

f. R

. Rom

ansk

y, D

. Sc.

T

echn

ical

Uni

vers

ity –

Sof

ia

8, K

limen

t Ohr

idsk

i Blv

d., 1

000

Sof

ia, B

ULG

AR

IA

E-m

ail:

rro

m@

tu-s

ofi

a.b

g

WE

B S

ite:

htt

p:/

/info

tech

-bg

.co

m

Tel

.: (

+35

9 2)

965

-32-

95 /

965-

25-3

0 F

ax (

offic

e):

(+35

9 2)

962

-45-

77

32

nd I

nte

rnati

onal C

onfe

rence

on I

nfo

rm

ati

on T

echnolo

gie

s

20

-21

Se

pte

mb

er

20

18

S

t. S

t. C

on

sta

nti

ne

an

d E

len

a r

es

ort

B

UL

GA

RIA

CA

LL F

OR

PA

PE

RS

(Fir

st

Annou

ncem

ent)

Pa

pe

rs i

nd

ex

ing

& a

bs

tra

cti

ng

in

:

Ne

w fo

rmat

of t

he e

vent

:

Phas

e (1

): Vi

rtua

l for

um &

pre

limin

ary

e-di

scus

sion

Phas

e (2

): On

site

pre

sent

atio

n of

acc

epte

d re

port

s

TO

PIC

S

We in

vite subm

issio

ns of

papers

pre

senting origin

al

re-

searc

hes in the f

ollo

win

g topic

s (

but not lim

ited):

In

form

ati

on

Tech

no

log

ies

Sof

twar

e T

echn

olog

ies

and

Pro

gram

min

g; D

igita

l S

igna

l P

roce

ssin

g an

d A

pplic

atio

ns;

Dat

abas

e M

anag

emen

t S

yste

ms

and

Info

rmat

ion

Sys

tem

s; B

usin

ess

Inte

llige

nce,

CR

M-

and

ER

P-S

yste

ms;

Pro

ject

M

anag

emen

t an

d Q

ualit

y M

anag

emen

t S

yste

ms;

Com

puta

tiona

l So-

cial

Sci

ence

; Hyp

erm

edia

, Mul

timed

ia a

nd S

ocia

l Med

ia; V

isua

lizat

ion

Met

hods

and

Tec

hniq

ues;

Aug

men

ted

and

Virt

ual

Rea

lity;

Bio

info

r-m

atic

s an

d H

ealth

Info

rmat

ics.

In

form

ati

on

Secu

rity

Sec

urity

Pol

icy,

Met

hods

and

Too

ls;

Info

rmat

ion

The

ory

& C

oddi

ng

The

ory;

Cry

ptog

raph

y, P

KI

and

e-S

igna

ture

; B

iom

etric

s an

d A

uthe

n-tic

atio

n T

echn

olog

ies;

Dig

ital R

ight

s M

anag

emen

t S

yste

ms;

Sec

urity

P

olic

y A

naly

sis

and

Cris

is M

anag

emen

t; C

yber

-att

acks

and

Sec

urity

; E

nerg

y P

olic

y an

d S

ecur

ity o

f Sup

plie

s.

N

etw

ork

ing

an

d C

om

mu

nic

ati

on

Tech

no

log

ies

Net

wor

king

– P

roto

cols

, Des

ign,

Inve

stig

atio

n; P

2P a

nd C

ompl

ex N

et-

wor

ks; G

rid S

yste

ms;

Clo

ud C

ompu

ting

(Clo

ud S

ervi

ces

and

App

lica-

tions

); M

obile

Clo

ud C

ompu

ting;

Int

erne

t of

Thi

ngs

(IoT

); M

achi

ne t

o M

achi

ne C

omm

unic

atio

ns (

M2M

); A

uton

omic

Net

wor

k C

ompu

ting;

O

nlin

e C

omm

uniti

es a

nd S

ocia

l Com

putin

g; W

eb-B

ased

App

licat

ions

.

In

tellig

en

t S

yste

ms a

nd

Ap

plicati

on

s

Inte

llige

nt a

nd A

gent

Sys

tem

s; K

now

ledg

e-B

ased

App

licat

ions

; O

n-to

logy

and

Sem

antic

Web

; In

telli

gent

Tra

nspo

rt S

yste

ms;

Int

elle

ctua

l an

d V

irtua

l Env

ironm

ents

; U

ser

Mod

ellin

g.

T

ech

no

log

ies f

or

Syste

m D

esig

n &

In

vesti

gati

on

Aut

omat

ion

of S

yste

m D

esig

n an

d R

esea

rch;

Com

pute

r A

rchi

tect

ure

and

Per

form

ance

Eva

luat

ion;

Par

alle

l Sys

tem

s an

d T

echn

olog

ies;

Mi-

croa

rchi

tect

ure

and

Rec

onfig

urab

le C

ompu

ting;

Com

pute

r M

odel

ling

and

Sys

tem

Inv

estig

atio

n; C

AD

Sys

tem

s; E

mbe

dded

Sys

tem

s; T

est-

ing,

Dia

gnos

tic a

nd S

yste

m R

elia

bilit

y; P

ower

Sys

tem

s –

Aut

omat

ion

and

Con

trol

; R

enew

able

Ene

rgy,

Sm

art

Sys

tem

s an

d E

nerg

y E

ffi-

cien

cy.

T

ech

no

log

ical A

sp

ects

of

e-G

overn

an

ce &

Pri

vacy

e

-Soc

iety

(e-

Pol

icy,

e-D

emoc

racy

, e-

Vot

ing)

; e-

Acc

ess

and

e-S

er-

vici

ng;

e-G

over

nmen

t an

d e-

Mun

icip

ality

; e-

Lear

ning

– S

yste

ms

and

App

licat

ions

; e-

Bus

ines

s an

d e-

Com

mer

ce;

e-H

ealth

and

Int

erop

era-

bilit

y in

Hea

lthca

re IC

T.

P

rivac

y an

d P

erso

nal D

ata

Pro

tect

ion

(PD

P);

Mod

els

of P

DP

, Reg

-is

trat

ion

and

Reg

ulat

ion;

Pol

icy

and

Pra

ctic

es f

or P

DP

; Tec

hnol

ogic

al

Str

uctu

res

for

PD

P; S

ecur

ity P

olic

y in

PD

P; P

DP

in C

loud

Ser

vice

s.

WO

RK

ING

LA

NG

UA

GE

Englis

h

APPLIC

AT

ION

FO

R P

AR

TIC

IPA

TIO

N

To

take

par

t in

the

Info

Tec

h-2

018

you

shou

ld s

end

by th

e e-

mai

l b

efo

re A

pri

l 30t

h, 2

018,

the

fol

low

ing:

A

pplic

atio

n F

orm

(en

clos

ed in

this

for

m);

F

ull-l

engt

h pa

per

(com

plet

e m

anus

crip

t in

Eng

lish)

The

pap

er s

houl

d be

pre

pare

d in

one

col

umn

form

at (

just

ified

) us

ing

Tim

es N

ew R

oman

14p

t typ

efac

e on

A4

(210

mm

x297

mm

) sh

eet w

ithou

t a p

agin

atio

n. M

argi

ns: t

op –

30m

m, b

otto

m, l

eft a

nd

right

– 2

0mm

. T

he g

uide

for

pap

er p

repa

ratio

n (T

EM

PLA

TE

) is

av

aila

ble

on th

e co

nfer

ence

web

site

.

The

pap

er m

ust

cont

ain:

th

e tit

le o

f the

rep

ort i

n ca

pita

ls (

18pt

, b

old

);

th

e na

me(

s) o

f au

thor

(s)

– 14

pt (

bold

), a

ddre

ss(e

s) a

nd

coun

try

(14p

t, ita

lic);

ab

stra

ct (

4-6

row

s, 1

2pt)

and

key

wor

ds;

th

e co

mpl

ete

text

(in

trod

uctio

n, m

etho

d, r

esul

ts,

orig

inal

-ity

, app

licat

ion,

con

clus

ion

and

refe

renc

es).

Ele

ctro

nic

vers

ions

of

th

e A

pplic

atio

n F

orm

an

d of

th

e T

EM

PL

AT

E fo

r pa

per

prep

arat

ion

coul

d be

dow

nloa

ded

from

the

conf

eren

ce w

eb s

ite:

htt

p:/

/info

tech

-bg

.co

m.

For

all

subm

is-

sion

s pl

ease

, us

e e-

mai

l add

ress

: rr

om

@tu

-so

fia.

bg

F

or m

ore

info

rmat

ion

abou

t co

nfer

ence

his

tory

, pr

oced

ures

and

ph

oto

galle

ry y

ou c

ould

vis

it th

e C

onfe

renc

e w

eb s

ite.

PU

BLIS

HIN

G

The

pap

ers

will

be

revi

ewed

for

the

ir sc

ient

ific

cont

ent

by t

he

mem

bers

of t

he I

nt’l

Pro

gram

Com

mitt

ee a

nd a

utho

rs w

ill b

e no

-tif

ied

(by

Sec

ond

For

m)

of a

ccep

tanc

e or

rej

ectio

n. A

ccep

tanc

e w

ill b

e ba

sed

on q

ualit

y, r

elev

ance

and

orig

inal

ity.

The

rev

ised

an

d fin

al v

ersi

ons

of a

ccep

ted

pape

rs m

ust

be s

ubm

itted

in

cam

era-

read

y fo

rm b

efor

e Ju

ne

30th

, 20

18.

All

acce

pted

pap

ers

with

par

ticip

atio

n fe

e pa

ymen

t w

ill b

e pu

b-lis

hed

in t

he v

irtua

l e-f

orum

on

the

conf

eren

ce w

eb s

ite f

or p

re-

limin

ary

disc

ussi

on (

15 J

uly

– 1

5 A

ug

ust

) an

d w

ill b

e in

clud

ed in

th

e C

onfe

renc

e P

rogr

am fo

r pre

sent

atio

n on

site

in re

port

/ po

ster

co

nfer

ence

ses

sion

s.

The

pap

ers

nom

inat

ed a

fter

the

on-

line

disc

ussi

on b

y po

sitiv

e co

nclu

sion

will

be

publ

ishe

d in

the

PR

OC

EE

DIN

GS

(IS

SN

131

4-10

23)

whi

ch w

ill b

e sp

read

ove

r 20

libr

arie

s in

all

over

the

Wor

ld.

The

ele

ctro

nic

vers

ion

of t

he p

ublic

atio

ns

will

be

incl

uded

in t

he

scie

ntifi

c da

taba

ses

of th

e E

BS

CO

Hos

t (U

SA

).

Inte

rnati

onal Confe

rence

Info

Tech-2

01

8

AP

PL

ICA

TIO

N F

OR

M

(ple

ase,

dow

nloa

d th

e fo

rm fr

om th

e co

nfer

ence

web

site

)

Part

icip

an

t:

Nam

e,

Surn

am

e:

Title

/Degre

e:

Org

aniz

ation:

Tel.:

E-m

ail:

Mail A

dd

ress (

for

po

sta

l m

ate

rials

):

Fo

rm o

f p

art

icip

ati

on

:

K

eynote

Sta

ndard

paper

S

hort

paper

Ple

nary

rep

ort

8-1

0 p

ages

4-5

pag

es

Topic

:

Title

of paper:

A

uth

or(

s):

D

ate

: . . . . . . . . . . .

S

ignatu

re:

. . . . . . . . . . .

Subm

issio

ns m

ust arr

ive b

efo

re A

pri

l 30

th,

20

18.

The fin

al

decis

ion a

bout th

e m

anner of pre

senta

tion o

f th

e p

apers

will

be t

aken b

y t

he P

rogra

m C

om

mitte

e.

THE FEDERATION OF THE SCIENTIFIC ENGINEERING UNIONS IN BULGARIA (FNTS)

FNTS is a professional, scientific-educational, non-governmental, non-political, non-profit association of legal entities – professional organizations registered

under the Law as non-profit legal entities, whose members are engineers, economists, and other specialists in the

fields of science, technology, economy, and agriculture.

FNTS has extensive bilateral cooperation with similar organizations in many countries.

FNTS brings together 19 national associations – Scientific and Technical Unions (STU) and 34 Territorial

associations, which have more than 15 000 highly educated professionals from across the country.

FNTS is a co-founder and active member of the World Federation of Engineering Organizations (WFEO).

FNTS is a member of the European Federation of National Engineering Associations (FEANI), and member of the Standing Conference of engineering organizations

from Southeast Europe (CO.PICEE), Global Compact, European Young Engineers (EYE). The Federation has the

exclusive right to award nominations for the European Engineer (EUR ING) title.

Contacts: 108 Rakovski St., Sofia 1000, Bulgaria

web: www.fnts.bgemail: [email protected]

ELECTROTECHNICA & ELECTRONICA E+E...2018/01/02 · ELECTROTECHNICA & ELECTRONICA E+E Vol. 53. No...

Documents

Transcript of ELECTROTECHNICA & ELECTRONICA E+E...2018/01/02 · ELECTROTECHNICA & ELECTRONICA E+E Vol. 53. No...