Ambient Intelligence (AmI) - 123seminarsonly.com...Ambient Intelligence (A mI) is a new paradigm in...

AAmmbbiieenntt IInntteelllliiggeennccee ((AAmmII))

SEMINAR REPORT2009-2011

In partial fulfillment of Requirements inDegree of Master of Technology

InCOMPUTER & INFORMATION SCIENCE

SUBMITTED BY

RENETHA J B

DEPARTMENT OF COMPUTER SCIENCECOCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY

KOCHI – 682 022

COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGYKOCHI – 682 022

DEPARTMENT OF COMPUTER SCIENCE

CCEERRTTIIFFIICCAATTEE

This is to certify that the seminar report entitled “AAmmbbiieenntt IInntteelllliiggeennccee ((AAmmII))”” is being

submitted by RReenneetthhaa JJ BB in partial fulfillment of the requirements for the award of

M.Tech in Computer & Information Science is a bonafide record of the seminar

presented by her during the academic year 2010.

Mr. G.Santhosh Kumar Prof. Dr.K.Poulose JacobLecturer DirectorDept. of Computer Science Dept. of Computer Science

AACCKKNNOOWWLLEEDDGGEEMMEENNTT

First of all let me thank our Director Prof: Dr. K. Poulose Jacob, Dept. of

Computer Science, CUSAT who provided with the necessary facilities and advice. I am

also thankful to Mr. G.Santhosh Kumar, Lecturer, Dept of Computer Science,

CUSAT for his valuable suggestions and support for the completion of this seminar.

With great pleasure I remember Dr. Sumam Mary Idicula, Reader, Dept. of Computer

Science, CUSAT for her sincere guidance. Also I am thankful to all of my teaching and

non-teaching staff in the department and my friends for extending their warm kindness

and help.

I would like to thank my parents without their blessings and support I would not

have been able to accomplish my goal. I also extend my thanks to all my well wishers.

Finally, I thank the almighty for giving the guidance and blessings.

ABSTRACT

Philips Research introduced Ambient Intelligence(AmI) in the year 1998. In 2001,

AmI was taken up by The European Commission’s Information Society Technologies

Advisory Group (ISTAG). In computing, AmI refers to electronic environments that are

sensitive and responsive to the presence of people. Ambient intelligence is a vision on the

future of consumer electronics, telecommunications and computing for the time frame

2010–2020.

The development of ambient intelligence applications that effectively adapt to the

needs of the users and environments requires the presence of planning mechanisms for

goal-oriented behavior. A planning system for AmI applications is based on the

hierarchical task network (HTN) approach and is called distributed hierarchical task

network (D-HTN). D-HTN is able to find courses of actions to address given goals.

The application areas of AmI include health-related applications, public

transportation sector, education services etc. This seminar aims to give an insight into

ambient intelligence technology and a planner for AmI applications.

Keywords: Ambient intelligence, context awareness, sensors, planning, multiiagents

CONTENTS

Slno Title Page no.

1 Introduction to AmI 1

2 History 2

3 AmI

3.1 Vision 3

3.2 Semantics 4

3.3 Key concepts 4

3.4 Key Technologies 5

4 Social and political aspects of AmI 6

5 Relation between AmI and other Computer Science areas 7

6 5Ws and 3Ps of AmI 9

7 Architecture of AmI system 11

8 Components of AmI system 12

9 AmI System - Planning

9.1 Features of AmI systems 15

9.2 Why planning needed for AmIApplications?

15

9.3 Planning and D-HTN planner 16

9.4 D-HTN algorithms 18

9.5 Application Scenario 19

10 Application areas 25

11 Challenges 26

12 Conclusion 28

13 References 29

Ambient Intelligence (AmI)

Dept. of Computer Science, Cochin University of Science & Technology 1

1. Introduction

Ambient Intelligence (AmI) is a new paradigm in Information Technology that has

potential for great impact in the future. The vision of AmI is that the people will be

surrounded by intelligent objects that can sense the context and respond according to the

desire of the people. AmI is a multidisciplinary topic, since it combines the features of many

of the areas in Computer Science.

In the last five years, we have seen significant advances in three promising technology

areas: virtual environments, in which 3D displays and interaction devices immerse the user in

a synthesized world, mobile communication and sensors, in which increasingly small and

inexpensive terminals and wireless networking allow users to roam the real world without

being limited to stationary machines. The merging of these areas allows the emergence of a

new vision: the Ambient Intelligence (AmI).

AmI refers to a digital environment that proactively, but sensibly, supports people in

their everyday lives. It will make the feeling that the people live with technology. It is aligned

with the concept of ‘disappearing computer’, since the AmI environment make the

technology invisible. As the devices grow smaller, more connected and more integrated into

our environment, the technology disappears into our surroundings.

“The most profound technologies are those that disappear. They weave themselves into thefabric of everyday life until they are indistinguishable from it.” M. Weiser

The basic idea behind AmI is that by enriching an environment with technology

(mainly sensors and devices interconnected through a network), a system can be built to take

decisions to benefit the users of that environment based on real-time information gathered

and historical data accumulated.

An important aspect of AmI has to do with interaction. On one side there is a

motivation to reduce the human-computer interaction as the system is supposed to use its

intelligence to infer situations and user needs from the recorded activities, as if a passive

human assistant was observing activities unfold with the expectation to help when (and only

if) required. On the other side, a diversity of users may need or voluntarily seek direct

interaction with the system to indicate preferences and needs. The entire environment around

us, homes and offices, cars and cities, will collectively develop a pervasive network of

intelligent devices that will cooperatively gather, process and transport information.



2. History

In 1998, the board of management of Philips commissioned a series of presentations

and internal workshops, organized by Eli Zelkha and Brian Epstein of Palo Alto Ventures

(who coined the name 'Ambient Intelligence') to investigate different scenarios that would

transform the high-volume consumer electronic industry from the current “fragmented with

features” world into a world in 2020 where user-friendly devices support ubiquitous

information, communication and entertainment. In the years after, these developments grew

more mature. In 1999, Philips joined the Oxygen alliance, an international consortium of

industrial partners within the context of the MIT Oxygen project, aimed at developing

technology for the computer of the 21st century. In 2000, plans were made to construct a

feasibility and usability facility dedicated to Ambient Intelligence. This HomeLab officially

opened on 24 April 2002.

Along with the development of the vision at Philips, a number of parallel initiatives

started to explore ambient intelligence in more detail. In 2001, the concept of Ambient

Intelligence (AmI) was taken up by European Commission’s Information Society

Technologies Advisory Group (ISTAG). The term Ambient Intelligence is defined by ISTAG

as "the convergence of ubiquitous computing, ubiquitous communication, and interfaces

adapting to the user". Following the advice of the ISTAG, the European Commission used

the vision for the launch of their sixth framework (FP5) in Information, Society and

Technology (IST), with a subsidiary budget of 3.7 billion euros. EU FP6: driving vision in a

3.7BEuro 5 year ICT (Information and Communication Technologies) research programme

(2002-2006). EU FP7 (9.1 BEuro for ICT): acknowledged (mainstreamed) but more

focussed, systemic and transformational (2007-2012)

The European Commission played a crucial role in the further development of the

AmI vision. As a result of many initiatives the AmI vision gained traction. Fraunhofer

Society started several activities in a variety of domains including multimedia, microsystems

design and augmented spaces. MIT started an AmI research group at their Media Lab.

Several more research projects started in a variety of countries such as USA, Canada, Spain,

France and the Netherlands. In 2004, the first European symposium on AmI (EUSAI) was

held and many other conferences have been held that address special topics in AmI.



3. AmI: Vision, semantics, key concepts and key technologies

3.1 Vision

Ambient Intelligence (AmI) will radically change how people interact with

technology. In AmI, people will be surrounded by a multitude of interconnected embedded

systems. These devices will be able to locate and recognize objects and people, as well as

people’s intentions.

The vision of AmI is characterized by two key features: intelligence and embedding.

The feature of “intelligence” refers to the fact that the digital environment is able to analyze

the context, adapt itself to the people and objects that reside in it, learn from their behavior,

and eventually recognize as well as express emotion. The feature of “embedding” means that

miniaturized devices will increasingly become part of the invisible background of peoples’

activities, and that social interaction and functionality will move to the foreground.

According to the AmI vision,” people will not just use technology: they will live with

it.”

Hence, AmI is :-

vision for our environment

‘smart electronic environments that are sensitive and responsive to the presence of

people’

‘Electronics embedded in every-day objects; natural interaction; context aware;

personalised; adaptive; responsive; pro-active.’

Enhancing productivity, healthcare, well-being, expressiveness, creativity.



3.2 Semantics

Ambient Intelligence refers to electronic environments that are sensitive and

responsive to the presence of people

The term ambient refers to the environment and reflects the need for typical

requirements such as distribution, ubiquity, and transparency.

Distribution refers to noncentral systems control and computation.

Ubiquity means the embedding is present everywhere.

Transparency indicates that the surrounding systems are invisible and

unobtrusive.

The term Intelligence means the digital surroundings exhibit specific forms of social

interaction. In other words, an environment must recognize the people that live in it, adapt

itself to them, learn from their behavior, and possibly show emotion. In short, the

environment should be intelligent.

3.3 Key Concepts

AmI provides ‘Smarter’ living. ie. AmI is a technology for people. To refine the

notion of ambient intelligence, Marzano and Emile Aarts formulated the following five key

concepts of AmI:

Embedded. Many networked devices are integrated into the environment.

Context aware. The system can recognize you and your situational context.

Personalized. The system can tailor itself to meet your needs.

Adaptive. It can change in response to you.

Anticipatory. The system anticipates your desires without conscious

mediation.

The first two elements relate to the integration of hardware devices into the

environment, and refer to embedded systems in general. Embedded systems play an



important role in the realization of ambient intelligence because they account for the

embedding of electronic devices into people’s surroundings.

The three other key elements of ambient intelligence concern the adjustment of

electronic systems in response to users. These system adjustments occur on different time

scales. Personalization refers to those occurring on a short time scale (for example, installing

personalized settings). Adaptation involves adjustments to changing user behaviors detected

by monitoring the user over longer periods of time. Ultimately, when the system gets to know

the user so well that it can detect behavioral patterns, adjustments are possible over a very

long period of time.

3.4 Key Technologies

The benefit of an AmI system is measured by how much can give to people while

minimizing explicit interaction. The aim is to enrich specific places (a room, a building, a car,

a street) with computing facilities which can react to people’s needs and provide assistance.

In order for AmI to become a reality a number of key technologies are required:

Unobtrusive hardware (Miniaturisation, Nanotechnology, smart devices,

sensors etc.)

Seamless mobile/fixed communication and computing infrastructure

(interoperability, wired and wireless networks, service-oriented architecture,

semantic web etc.)

Dynamic and massively distributed device networks, which are easy to control

and program (e.g. service discovery, auto-configuration, end-user

programmable devices and systems etc.).

Human-centric computer interfaces (intelligent agents, multimodal interaction,

context awareness etc.)

Dependable and secure systems and devices (self-testing and self repairing

software, privacy ensuring technology etc.)



4. The social and political aspects of ambient intelligence

Ambient intelligence is more than just a question of embedding technology into objects.

It involves human culture in its broadest sense: universal desires; complex social

relationships; diverse value systems; individual likes and dislikes; the sustainability of

economic and natural ecosystems; and codes of ethics, conduct, and communication, both in

civil society and in business. This is also what makes ambient intelligence markedly different

from other concepts such as pervasive computing and ubiquitous computing

In AmI, technology lives with the people, hence AmI has both social and political

influences. The current phase of AmI/pervasive computing, in which computers are already

being embedded in many devices, has begun to affect our everyday lives in ways we do not

even notice.

ISTAG identified a series of necessary characteristics that will permit the eventual

societal acceptance of AmI.

AmI should:

facilitate human contact.

be orientated towards community and cultural enhancement.

help to build knowledge and skills for work, better quality of work, citizenship and

consumer choice.

inspire trust and confidence.

be consistent with long term sustainability - personal, societal and environmental -

and with life-long learning.

be made easy to live with and controllable by ordinary people.



5. Relation between AmI and other Computer Science areas

Fig 1

Networks, Sensors, Human Computer Interfaces (HCI), Pervasive Ubiquitous

Computing and Artificial Intelligence (AI) are all relevant and interrelated but none of them

conceptually covers the full scope of AmI. Ambient Intelligence puts together all these

resources to provide flexible and intelligent services to users acting in their environments.

Ambient intelligence involves the convergence of several computing areas. It is a

multi-disciplinary approach which aims to enhance the way environments and people interact

with each other. The ultimate goal of the area is to make the places we live and work in more

beneficial to us.

First, a machine was shared by many highly trained programmers. Then it became

possible in many countries around the world that many people, not necessarily with a high



level of training, will have access to one PC in an individual basis. Now many people can

have access to several computing devices like a PC, a laptop and a PDA at work plus a PC at

home and various smaller processing units embedded in electro-domestic appliances.

All seems to indicate this trend will continue. Slowly systems are being designed in

such a way that people do not need to be a computer specialist to benefit from computing

power. This technical possibility is being explored in an area called Ambient Intelligence

(AmI) where the idea of making computing available to people in a non-intrusive way is at

the core of its values. The benefit of an AmI system is measured by how much can give to

people whilst minimizing explicit interaction. The aim is to enrich specific places (a room, a

building, a car, a street) with computing facilities

Fig 2



6. 5Ws and 3Ps of AmI

Of Importance for AmI are the “5Ws” (Who, Where, What, When and Why) principle

of design:

Who: the identification of a user of the system and the role that user plays within the system

in relation to other users. This can be extended to identifying important elements like pets,

robots and objects of interest within the environment.

Where: the tracking of the location where a user or an object is geographically located at

each moment during the system operation. This can demand a mix of technologies, for

example technology that may work well indoors may be useless outdoors and vice-versa.

When: the association of activities with time is required to build a realistic picture of a

system’s dynamic. For example, users, pets and robots living in a house will change location

often change location and knowing when those changes happened and for how long they

lasted are fundamental to the understanding of how an environment is evolving.

What: the recognition of activities and tasks users are performing is fundamental in order to

provide appropriate help if required. The multiplicity of possible scenarios that can follow an

action makes this very difficult. Spatial and temporal awareness help to achieve task

awareness.

Why: the capability to infer and understand intentions and goals behind activities is one of

the hardest challenges in the area but a fundamental one which allows the system to

anticipate needs and serve users in a sensible way

There seems to be a growing consensus that achieving sustainability requires a good

balance between three factors, sometimes referred to as the three P’s: people, planet, and

profit.

People: Humans exploit everything around them to improve their lives and expand their

powers. They want to acquire everything with minimum effort and maximum comfort. This

desire, to have devices that amplify human powers without hindering or cluttering their lives

is what drives the increasing miniaturization of devices. Many devices have already made the



transition from big static objects to small objects that people can carry around on their bodies.

Clocks are now wristwatches, and more recently phones and audio systems have reached the

stage of becoming worn on the body. This instinct to find greater comfort, power, knowledge,

and freedom has been the main driving force behind technological innovation.

Ambient intelligence intends to improve the quality of people’s lives. Not everything

that’s possible with technology is actually desirable. Therefore, it’s crucial that people make

the right choices with ambient intelligence. This is only possible if people agree on what

quality of life and what sort of world they would like to see develop.

Planet: AmI has a great contribution to the planet. AmI provides better care for the

environment. Numerous novel ecological developments are possible by integrating smart

electronics into the environment. They aid in checking pollution and checking uncontrolled

dumping of waste products. There are also techniques for determining energy wastage and

reduce needless consumption.

Profit: Ambient Intelligence describes a new economy called “experience economy”. It is

positioned as the fourth major wave following the classic economies of commodity, goods,

and service. People are willing to spend money for getting better experience. Recollection of

a personal event might just bring back that good old feeling.

Virtual worlds in an ambient-intelligent environment might support such events.

There are many other applications, such as ambient lighting, ambient sounds and poetic

interfaces which all could bring good feel to people. A salient property of an experience is

that it can feel real, whether it has been generated by a real or a virtual cause; what counts is

the belly feeling.



7. Architecture of AmI system

Fig 3

Sensors bring data to the system. The data collected is transmitted by the network and

pre-processed by the middleware, which collates and harmonises data from different devices.

In order to make decision-making easier and more beneficial to the occupants of the

environment the system will have a higher level layer of reasoning which will accomplish

diagnosis and advise or assist humans with responsibility for intervention.

Elements that may be included in the high level ‘Decision Making’ process are a

‘Knowledge Repository’ where the events are collected and an ‘AI Reasoner’ which will

apply for example spatio-temporal reasoning to take decisions. For example, a decision could

be to perform some action in the environment and this is enabled via ‘Actuators’. Knowledge

discovery and machine learning techniques learn from the acquired information in order to

update the AI Reasoner in the light of experience of the system.



8. Components of AmI system

AmI system is comprised of three main components: ubiquitous computing,

ubiquitous communication, and user adaptive interfaces.

Ubiquitous computing means any computing device, while moving with you, can

build incrementally dynamic models of its various environments and configure its services

accordingly. The devices will be able to either "remember" past environments they operated

in, or proactively build up services in new environments. Ubiquitous computing" refers to

omnipresent computers that serve people in their everyday lives at home and at work,

functioning invisibly and unobtrusively in the background and freeing people to a large

extent from tedious routine tasks. This includes pen-based technology, hand-held or portable

devices, large-scale interactive screens, wireless networking infrastructure, and voice or

vision technology.

Ubiquitous communication: Ubiquitous computing is the introduction and

expansion of wireless network technology, which enables flexible communication between

interlinked devices that can be stationed in various locations or can even be portable.

Wireless LAN (W-LAN) applications per standard IEEE 802.11b offer high-speed

transfer rates of 11 Mbit/s and can be extended over entire office buildings and production

areas by using several access points. While W-LAN is considerably cheaper than a traditional

stationary LAN, it is often still too costly to be included in small individual devices

Bluetooth technology is used in today's handheld applications like cellular phones or

personal digital assistants (PDAs) per standard IEEE 802.15 to allow wireless connection

within a personal area network (W-PAN). While the cost of Bluetooth equipment is

significantly lower than the cost of W-LAN, the transmission range of up to 10 meters and

the data transfer rate of less than 720 Kbit/s are inferior. New Bluetooth versions are

currently under development that attempt to eliminate the latter drawback. V1.2 allows rates

of up to 3 Mbit/s, V2.0 of up to 12 Mbit/s

High rate W-PANs per standard IEEE 802.15 TG3, launched in 2003, use higher

power devices (8 dBm) than regular Bluetooth equipment (0 dBm) to transmit data at a rate

of up to 55 Mbit/s and over a range of up to 55 m. This technology is, therefore, an attractive



alternative to W-LAN, especially considering the comparatively lower cost.

Low power W-PANs per standard IEEE 802.15 TG4 are particularly useful for handheld

devices since energy consumption for data transmission purposes, and costs, are extremely

low. The range of operation of up to 75 m is higher than current Bluetooth applications, but

the data transfer rate of 250 Kbit/s is lower.

Wireless body area networks (BANs) interlink various wearable devices, such as

wireless data glasses, earpieces, microphones, and sensors, and can connect them to outside

networks. BANs are often used for medical applications but also in work-related fields, for

example, to provide production operators with instructions that are adapted to the respective

work situation. BANs usually consist of a central network unit, which connects the devices

and which can provide an interface to further networks outside the BAN, for example, via

Bluetooth. Advantages of BANs versus W-PANs are the short range and the resulting lower

risk of tapping and interference, as well as low frequency operation, which leads to lower

system complexity. Technologies used for wireless BANs include magnetic, capacitive, low-

power far-field and infrared connections, while non-wireless BANs use wires or conductive

fabrics.

Radio frequency identification (RFID) encompasses wireless identification through radio

transmission. RFID systems comprise a read/write station and active (with own power

source) or passive (power supplied by the read/write station) transponders (transmitter /

responder), and can be used in a variety of applications. Traditional examples include

protection against theft, access control, and billing. The range of possible applications is

much greater: RFID systems can be used for material tracking in manufacturing and logistics,

for cash register applications in stores as an alternative to barcode scanning, or for localizing

items or persons.

Network administration is facilitated by minimizing the effort required for setting up

networks. The introduction of mobile ad hoc networks (MANETs) is an important step in this

direction. A MANET uses the wireless technologies described in the list above but is more

flexible than conventional networks, since the routers are included in the mobile nodes

instead of being fixed and have the ability to configure themselves. This provides the network

with great flexibility due to its ability to adapt automatically to a changing network

environment.



User adaptive interfaces

User adaptive interfaces, the third integral part of AmI, are also referred to as

"Intelligent social user interfaces" (ISUIs). These interfaces go beyond the traditional

keyboard and mouse to improve human interaction with technology by making it more

intuitive, efficient, and secure. They allow the computer to know and sense far more about a

person, the situation the person is in, the environment, and related objects than traditional

interfaces can.

ISUIs encompass interfaces that create a perceptive computer environment rather than

one that relies solely on active and comprehensive user input. ISUIs can be grouped into five

categories:

Visual recognition (e.g. face, 3D gesture, and location) and output

Sound recognition (e.g. speech, melody) and output

Scent recognition and output

Tactile recognition and output

Other sensor technologies

The key to an ISUI is the ease of use, in this case the ability to personalize and

adapt automatically to particular user behavior patterns (profiling) and different situations

(context awareness) by means of intelligent algorithms. In many cases, different ISUIs, such

as voice recognition and touch screen, are combined to form multi-modal interfaces. ISUIs

make network usage more secure as the interfaces can identify users automatically by, for

example, face or voice recognition instead of requesting a password.



9. AmI system - Planning

9.1 Features of AmI Systems

AmI system is composed of numerous agents. Agents are smart devices, which are

fixed or mobile devices. Agents form part of AmI system either permanently or temporarily.

For example a person comes with a mobile phone into a room equipped with AmI system.

The cell phone, when properly connected to the network of other devices, is temporarily part

of the system. After the person leaves the room is disconnected.

Features of AmI system are:-

• Feature 1: Some agents could take no responsibility in building the plan because of their

limitations in processing and communication. This pushes toward the centralized

planning process.

• Feature 2: The skills to perceive the environment and to perform the actions are

distributed over the agents. This pulls toward the distributed planning process.

9.2 Why Planning needed for AmI applications?

The development of ambient intelligence (AmI) applications that effectively adapt to

the needs of the users and environments requires the presence of planning mechanisms for

goal-oriented behavior. An AmI system that plans is able to find a course of action that, when

executed, achieves a desired effect. The planning system builds plans according to the

capabilities of available devices that perform actions to satisfy the user’s need.

A planning system for AmI applications proposed by Francesco Amigoni, Associate

Member, IEEE and Nicola Gatti, Member, IEEE, is based on the hierarchical task network

(HTN) approach and it is called distributed hierarchical task network (D-HTN). D-HTN

planner can support both the features of AmI systems; i.e centralized as well as distributed

features.



9.3 Planning and D-HTN planner

A planning algorithm has three inputs:

– a description of the world,

– a description of the goal, and

– a description of the capabilities in form of possible actions that can be

performed.

The planning algorithm’s output is a sequence of actions such that, when they are executed in

a domain satisfying the initial state description, the goal will be achieved. AmI system need a

centralized planner that manages distributed capabilities.A distributed HTN approach

appears appropriate for AmI applications because it naturally supports heterogeneous agents

and knowledge exchange among them.

D-HTN planners are based on the concept of task network that is represented as

[(n1:1 ),(n2:2 ),……(nm: m), ]

where

i are tasks, either primitive (that can be directly executed by an agent) or

nonprimitive (that must be further decomposed);

ni are labels to distinguish different occurrences of the same task;

is a Boolean formula representing the constraints on the tasks, such as variable

bindings constraints [e.g.,v=v’], ordering constraints [e.g., (n<n’), with the meaning

that n must be executed before n’], and state constraints [e.g.,(n,l,n’) , with the

meaning that l must be true immediately after n, immediately before n’, and in all

states between n and n’ ].

A task network can be represented by a graph. For example, the task network:



Fig 4The intended meaning of this graph is that, in order to request a good g1 by e-mail, we

first have to create the RequestText t1 and look for the EmailAddress a1 of a supplier of g1,

and then we have to SendEmail with content t1 to a1.

Functions of Agents and Planner in D-HTN planner:-

• AGENT:

– Each agent keeps a local data structure called plan library, which stores all thedecompositions it knows.

– The decompositions in the plan library of an agent have been defined by thedesigner during the installation of the agent and are peculiar for each agent

• PLANNER:

– generate a plan, the other agents are only requested to communicatedecompositions .

By means of a communication mechanism based on message passing,

– the planner can ask the currently connected agents to send their availabledecompositions for a given task

– the agents can send to the planner the requested decompositions.

D-HTN planning starts with an initial task network D representing the problem (the

goal) and with a set M of methods or decompositions. Each decomposition is a pair

m=(t,d),where t is a nonprimitive task and d is a task network; m says that a way to achieve is

to perform the tasks in . Then, D-HTN planning proceeds by finding a nonprimitive task from



the current task network D and a method m=(t’,d’), in M such that t’ unifies with t and by

replacing t with d’ in D. When only primitive tasks are left in D, a plan for the original

problem can be found. A plan is a sequence of ground primitive tasks .This pure HTN

planning process can be refined to make it more efficient by introducing backtracking, critic

functions, and other technicalities.

Each decomposition has associated three numerical indexes that are associated to:-

– Performance -measures the expected effectiveness of the decomposition

– Cost- measures the expected resource consumption for performing the tasks in

the decomposition

– probability of success - measures the expected likeliness that no error occurs

9.4 D-HTN Algorithms

D-HTN is composed of a set of distributed algorithms that are executed concurrently

by the planner and by the agents. Algorithm 1 presents an overview of the D-HTN algorithm

executed by the planning agent. The main data structure to represent the plan that is being

formed is a task network D. D is initialized with the initial task to be solved (i.e., the goal to

be reached). The D-HTN planner produces a final plan D composed only of primitive tasks

that can be executed by the agents. M(t) denotes the decomposition set.

Algorithm 1 D-HTN algorithm for the planner

1: D = initial task

2: while D contains nonprimitive tasks do

a) choose a nonprimitive task t from D

b) populate M(t), by requesting the currently connected agents to send

the decompositions m = (t’, d’) such that t’ unifies with t and by

collecting these decompositions

c) choose a decomposition m = (t’,d’) from M(t)

d) if t is primitive for the agent a proposing m then

bind a to t and remove t from the nonprimitive tasks

e) end iff) replace t with d’ in D

3:end while



Algorithm 2 D-HTN algorithm for the agents

1: while the agent is active do

a) wait for a message from the planner

b) if the message is a request of decompositions for a nonprimitive task t

then

send to the planner the decompositions m = (t’, d’) in the plan

library such that t’ unifies with t

c) end if

2: end while

9.5 Application Scenario

Consider a diabetic patient equipped with on-body monitoring devices who is

in a room equipped with AmI devices. In the diabetes case, the patient seldom

becomes suddenly ill because the monitoring devices can usually detect potential

alarming situations some time before they appear. AmI system is shown in fig 5.

Fig 5



The agent that simulates the monitoring devices on the diabetic patient is called goal

generator agent because in our application scenario it is the source of goals that the planning

system attempts to achieve. The goal generator agent stands for any device or user that can

generate a goal for the AmI system. The goal generator agent provides the input to our

planner in terms of high-level goals to be reached. The agents that populate the room in our

scenario are conceptually organized in three main classes: communication agents, repository

agents, and interactive agents.

The communication agents include the SMS agent, the email agent, the fax agent, and

the phone agent for sending and receiving SMS, e-mails, faxes, and phone calls, respectively.

The repository agents are the address book agent, a database of contacts, and the medical

store agent, a database of medicines currently present in the environment. The interactive

agents provide the sensors and the actuators to interact with the environment; they include the

thermometer agent, a temperature sensor, and the heating agent that can change the

temperature in the environment. All these agents (the goal generator agent and those

equipping the room) are supervised and coordinated by the environmental majordomo agent.

The agents represent devices that are physically and permanently part of the room (e.g.,

the heating agent) and the agents that represent mobile devices that are transiently part of the

room (e.g., the phone agent could be a cell phone carried by a person walking through the

room).

Some of the decompositions initially included in the plan libraries of some of the agents

composing the AmI system for the application scenario are listed in Fig 6.



Fig 6

Example :-

Consider a planner in solving the goal CheckAndRequest (Insulin) that is intended

to check the presence of insulin in the medical store and, if no insulin is left, to make a

request to pharmacies to provide insulin. The environmental majordomo asks the agents

currently connected to the AmI system to send their available decompositions.

The only decomposition for CheckAndRequest is provided by the environmental

majordomo itself and introduces two nonprimitive tasks, IsThere(Insulin) and

Request(Insulin).



Fig 7

The two tasks are connected by a selection statement: if the output of the execution of

the task IsThere (Insulin) (that checks if some insulin is left in the room) is False then the task

Request (Insulin) (that requests to supply insulin) is executed. The planning process picks up

the nonprimitive task IsThere (Insulin). This task can be decomposed only by a primitive task

of the medical store agent.

Fig 8Then, the nonprimitive task Request (Insulin) is considered. The SMS, e-mail, fax,

and phone agents propose their decompositions for the task in order to contact the pharmacies

by different communication means. Since the performance index value of the decomposition

proposed by the SMS agent is the highest, this is selected.



It is composed of nonprimitive tasks SearchCellNumber(Insulin,c1), SendSMS(t1,.c1)and CreateRequestText(Insulin,t1)

Fig 9

The selected decomposition constraints the tasks CreateRequestText(Insulin,t1) and

SearchCellNumber(Insulin,c1) to be executed before SendSMS(t1,.c1). The first task creates

the text containing t1 the request to supply insulin. The second task finds the mobile phone

numbers of the pharmacies. The SendSMS task sends out the requests.

The planning process continues and the new inserted nonprimitive tasks are

decomposed in primitive tasks performed by the goal generator, the address book agent, and

the SMS agent.



The planning process picks up the nonprimitive task CreateRequestText(Insulin,t1).

This task can be decomposed only by a primitive task of the goal generator agent. Next

nonprmitive task SearchCellNumber(Insulin,c1) can be decomposed only by a primitive task

of the address book agent. . Next nonprmitive task SendSMS(t1,c1) can be decomposed only

by a primitive task of the SMS agent.

Fig 10

The plan is now complete and ready to be executed. The execution of the plan is

supervised by the environmental majordomo that requests the agents to perform the primitive

tasks they proposed. During the execution, first the medical store is checked for insulin; if it

is found its counter is decremented by 1 and the plan execution ends since the task IsThere

(Insulin) returns True; otherwise, the execution continues by activating the other agents.



10. Application areas

Ambient Intelligence possesses applications in many areas. Some of them are listed below:-

• Health-related applications. Hospitals can increase the efficiency of their services by

monitoring patients’ health and progress by performing automatic analysis of activities in

their rooms. They can also increase safety by, for example, only allowing authorized

personnel and patients to have access to specific areas and devices.

• Public transportation sector. Public transport can benefit from extra technology including

satellite services, GPS-based spatial location, vehicle identification, image processing and

other technologies to make transport more fluent and hence more efficient and safe.

• Education services. Education-related institutions may use technology to track students

progression on their tasks, frequency of attendance to specific places and health related issues

like advising on their diet regarding their habits and the class of intakes they opted for.

• Emergency services. Safety-related services like fire brigades can improve the reaction to a

hazard by locating the place more efficiently and also by preparing the way to reach the place

in connection with street services. The prison service can also quickly locate a place where a

hazard is occurring or is likely to occur and prepare better access to it for security personnel.

• Production-oriented places. Production-centred places like factories can self-organize

according to the production/demand ratio of the goods produced. This will demand careful

correlation between the collection of data through sensors within the different sections of the

production line and the pool of demands via a diagnostic system which can advice the people

in charge of the system at a decision-making level



11. Challenges

Fig 11

The fast penetration of wireless communications has put into evidence the user’s need

to get easily connected anywhere and anytime at an affordable price. On the one hand,

wireless communications clearly proved that the most a technology provides simple access

means, added to freedom of movement and increased security, the most the user is willing to

accept it.

On the other hand, the most a technology is complex and costly, the less the user is

prone to accept it, in spite of possibly large potential advantages, which are generally not

reachable by the average user not interested in spending time and energies in acquiring the

underlying technology fundamentals. As a consequence, the successful systems of the future

will adhere to the paradigm of ”disappearing technologies”, both valid for communications

and computing, and will provide improved ease-of use at the expense of an increased, but

invisible to the user, complexity of the underlying systems and networks necessary to

transport and process the information in the different multimedia forms and usage contexts.



Ambient Intelligence faces a lot of challenges. Among these are the social implications

of AmI environments, the different potentials of AmI to enrich our lives, aspects of privacy

and trust, characteristics of different AmI interactions, how much intelligence people are

willing to accept, the different dimensions of the term ambient, the design of future

interaction spaces and intelligent artifacts, factors of user experience for implicit interaction,

existing and emerging AmI application areas and scenarios, the connection of AmI concepts

to physical spaces where it happens etc.

• Challenges in Interaction technology

Develop ambient interaction concepts that are truly intelligent,simple, and

intuitive.

Integrate multi-modality with context awareness and intuitive feedback

mechanisms.

Integrate smart media access into surroundings (audio, video, and light).

Develop interaction concepts for novel AmI technologies (photonictextiles, e-

paper, polymer lighting, and uld’s)

• Challenges in Innovation

Build an eco-system that uses co-creation as a model for open innovation.

Involve multiple parties in the user centered design cycle at large.

Concentrate on well-defined business domains (i.e., hospitality, fashion,

furniture, well-being, city beautification).

Develop new business models for AmI innovation

• Challenges in Involvement

Reach out to ordinary people so as to let them participate in the AmI effort.

Involve ordinary people in the user centered design cycle at large.

Let people experience the AmI future and live in it yourselves.

Make AmI part of education.



12. Conclusion

Ambient Intelligence (AmI) is growing fast as a multi-disciplinary topic of interest

which can allow many areas of research to have a significant beneficial influence into our

society. AmI is a vision on the future of consumer electronics, telecommunications and

computing for the time frame 2010–2020.

Ambient Intelligence envisions a world where people are surrounded by intelligent

and intuitive interfaces embedded in the everyday objects & physical environments around

them. These interfaces recognize and respond to the presence and behaviors of an individual

in a personalized and relevant way.

The new paradigm of ambient intelligence can bring about a revolution in the design,

appearance, and use of electronics in ordinary life. It could support and facilitate simple and

recurrent tasks, but it could also lead to a culture very different from today’s. This new

culture could develop through the expansion of the use media into a world in which physical

and virtual experiences merge to support personal expression, business productivity, and

personal lifestyles

Technology will not be the limiting factor in realizing ambient intelligence. The

ingredients to let the computer disappear are already available, but the true success of the

paradigm will depend on the ability to develop concepts that allow natural interaction with

digital environments.



13. References

[1] Francesco Amigoni, Associate Member, IEEE, Nicola Gatti, Member, IEEE, Carlo

Pinciroli, and Manuel Roveri, “What Planner for Ambient Intelligence

Applications?” ,IEEE Transactions On Systems, Man, And Cybernetics—Part A:

Systems And Humans , Vol. 35, No. 1, January 2005

[2] Emile Aarts, Philips Research, “Ambient Intelligence: A Multimedia Perspective”,

Published by the IEEE Computer Society, January–March 2004

[3] Carlos Ramos, Polytechnic of Porto • Juan Carlos Augusto, University of Ulster Daniel

Shapiro, Institute for the Study of Learning and Expertise, “Ambient Intelligence—the

Next Step for Artificial Intelligence”, Published by the IEEE Computer Society,

March/April 2008.

[4] Philips Research technology magazine , Password: Issue 23 • May 2005

[5] Juan Carlos Augusto and Paul McCullagh School of Computer Science and Mathematics

University of Ulster at Jordanstown BT37 0QB United Kingdom, “Ambient

Intelligence: Concepts and Applications ”

[6] Nigel Shadbolt ,University of Southampton, “Ambient Intelligence”, Published by the

IEEE Computer Society, July/August 2003

Ambient Intelligence (AmI) - 123seminarsonly.com...Ambient Intelligence (A mI) is a new paradigm in...

Documents

Transcript of Ambient Intelligence (AmI) - 123seminarsonly.com...Ambient Intelligence (A mI) is a new paradigm in...