IMPLEMENTATION OF ADAPTIVE MODULATION FOR A/G COMMUNICATION SYSTEM USING ZEPTOSDR

Zakaria El Alaoui Ismaili and Wessam Ajib, CoFaMic Research Center, Computer Sciences Department, Université du Québec à Montréal, Montréal, Canada

Omar A. Yeste-Ojeda and René Landry, Jr., LASSENA Research laboratory, Electrical Engineering Department, École de Technologie Supérieure, Montréal, Canada

Abstract

Software Defined Radio (SDR) technology is gaining popularity because of its capability to implement different applications with almost no hardware changes. Such a solution can decrease size, weight, power and cost requirements of avionics systems. This paper presents an experimental implementation of adaptive modulation technique for wideband air-to-ground (A/G) communication systems using Nutaq’s ZeptoSDR software defined radio platform. The system implementation provides a transmission of a real time captured video equipped by an adaptive modulation technique. The system autonomously chooses the most suitable modulation among BPSK, QPSK, 8PSK, 16QAM and 64QAM based on the estimation of link quality measured by the signal to noise ratio (SNR) in real time.

Keywords Adaptive modulation; Air-to-Ground avionic

communication; Software Defined Radio; SNR estimation.

I. Introduction Air-to-ground (A/G) communication systems

have been an interesting research topic for several years. Robustness, accuracy and high speed operational requirements attained by current avionics systems are far below what modern technology offers in many other fields of application. Future air-to-ground communications shall provide high degree of security and high communication capacity to provide voice, video, data transmission and internet service with the lower equipment cost. Nowadays, air-to-ground communication systems need higher transmission rates to ensure the various multimedia data services for commercial aircraft passengers. In

addition, the increase of airborne users requires solutions for more capacity and thus more spectral efficiency. Using high spectrally efficient modulations (four bits per symbol and more) should be a solution to increase the data rate. However, such modulations are more susceptible to noise due to aircraft speed, multipath problems and weather conditions. On the other hand, low spectrally efficient modulations (less than four bits per symbol) offer robust data transmission link rather less data rate. Therefore, a trade-off between high speed data transmission and received signal quality is required.

To overcome this problem, adaptive modulation technique provides a communications system with the ability to keep the bit error rate (BER) upper-bounded (almost constant) while maximizing the throughput in the channel. In good transmitting conditions, the system chooses the modulation offering maximum data rate, and when the channel conditions deteriorate, the system adapts to the situation by switching the modulation to a proper one which offers less data rate and a robust transmission.

Adaptive modulation is recognized as a basic technique that makes communication systems more reliable with as much transmission rate as possible. The approach of adapting modulations has been investigated and implemented in [1] for underwater acoustic communications to improve the system efficiency. The system provides several modulations schemes equipped by orthogonal frequency division multiplexing (OFDM). This technique is also used in satellite communications applications [2-4]. The system proposed in [2] analyzes adaptive modulation with 4QAM and 16QAM schemes whereas the system presented in [3] provides 8PSK and 16APSK modulations. Good performance is achieved using QPSK, 8PSK and 16APSK modulations in [4]. The adaptive modulation technique is also solicited in wireless access system applications [5-7]. On the

978-1-4799-8940-9/15/$31.00 ©2015 IEEE 2C5-1

other hand, there is a lack of use of the adaptive modulation in avionics applications.

Consequently, this paper presents a global overview of (A/G) links problematic followed by a quantitative study of adaptive modulation benefits for (A/G) communications systems. Then, an experimental implementation of adaptive modulation technique for (A/G) communication system using Nutaq’s ZeptoSDR software defined radio platform [8] is used for performance evaluation.

The system implementation provides a transmission of a real time captured video equipped by an adaptive modulation technique. The system autonomously chooses the most suitable modulation among BPSK, QPSK, 8PSK, 16QAM and 64QAM based on the estimation of link quality measured by the signal to noise ratio (SNR) in real time. Results are presented as comparison between the bit error rate performance theoretical curves and experimental ones to show how test results are close to theory models. Furthermore, results demonstrate how adaptive modulation can be beneficial to keep data transmission safe and robust regardless of the link constraints.

The rest of the paper is organized as follows. The air to ground link is discussed in Section II,

followed in Section III by a quantitative study and implementation of adaptive modulation. Test and performances are then presented and commented in Section IV and are followed by the conclusion.

II. Air-to-Ground Links Problematic Air-to-ground (A/G) communication systems

represent a powerful solution to provide broadband services for aircrafts. Such communications were proposed in order to increase air safety by ensuring voice and data services that are essential in aeronautical communications. These services can be roughly grouped into two categories:

• Safety related services represented by Air Traffic Services Control (ATSC) for pilots and controllers, and Air Operation Control (AOC).

• Non safety-related services which can be also provided in cabin for airline administrative Communications (AAC) or for passengers (APC).

Aeronautical (A/G) communication services are presented in Figure 1 [9].

Figure 1. (A/G) Communication Services

The (A/G) communication systems use the line-of-sight (LOS) in very high frequency (VHF) aeronautical narrowband (from 118 to 137 MHz with 25 kHz or 8.33 kHz per channel) whereas satellite communication (SATCOM) is solicited to ensure the links in oceanic areas [9]. However, the number of

aircrafts in the airspace is permanently increased which leads to VHF spectrum congestion. Therefore, Future Communications Infrastructure (FCI) and International Civil Aviation Organization (ICAO) recommended the use of L-band (between 960 ̶ 1164 MHz) termed L-band digital aeronautical

2C5-2

communication system (LDACS) [10] to decrease the spectrum saturation.

Cognitive radio was also proposed in [11] to be adopted in aeronautical A/G communications to provide broadband services for AOC and APC. Nevertheless, the VHF, L-band, low Earth orbit (LEO) SATCOM still the most promising candidates for A/G communication systems thanks to their high technical performance, low cost, and low risk [12].

Robustness and safety are fundamental factors for future (A/G) communication. Therefore, (A/G) systems shall provide a high degree of protection against frequency interferences which represent the main limitation in (A/G) communication.

Furthermore, continuous communication is always required. Accordingly, reliability and channel availability should be guaranteed as well in order to provide transmitted data integrity.

On the other hand, the increase of aircraft users requires solutions for more capacity (data rate) and thus more spectral efficiency to ensure more services. To increase the data rate, it is possible to use high spectrally efficient modulations. However, such modulations are more susceptible to noise. Meanwhile, low spectrally efficient modulations offer robust connexions rather less data rate. One solution can be efficient spectrum usage. Depending on the number of aircrafts in airspace and the throughput requirements, the transmission bitrate used by every aircraft can be dynamically adjusted.

For these reasons, adaptive modulation represents a key to bandwidth-efficient (A/G) communication by providing robust communication link with maximum throughput while keeping upper-bounded bit error rate. The adaptive modulation (AM) implementation for (A/G) communication systems will be presented in the next chapter.

III. (AM) Implementation

Hardware and Software Tools Nutaq’s ZeptoSDR is a fitting hardware

platform for adaptive modulation implementation [8].

This platform consists of Zynq-based Zedboard and Nutaq’s Radio420S FMC card providing good performance and cost trade-off. The Zedboard contains a coupled dual core ARM Cortex-A9 and the Artix-7 series FPGA. ZeptoSDR is able to guarantee both external and embedded host advanced programming interfaces (APIs).

Furthermore, the platform provides frequency agility thanks to tunable RF interface from 300 MHz to 3.8 GHz with the possibility to select the bandwidth from 1.5 to 28 MHz.

ZeptoSDR supports GNU Radio software suite to create complex SDR systems [13]. GNU Radio is an open source software development toolkit able to provide signal processing blocks to implement SDRs using off-the-shelf RF hardware. For GNU Radio applications, Python language is always used to write programs. In addition, C++ is solicited for performance-critical signal processing tasks. For running GNU Radio, Linux based operating system is compatible [14].

Analytical Principle Modulation agility depending on transmission

channel conditions represents a suitable solution for (A/G) systems bandwidth efficiency. While transmitting in good conditions, the system uses the modulation providing the maximum data rate. However, the system switches towards another modulation with less data rate to keep link robustness when the channel condition degrades.

This section presents the implementation of real-time captured video transmission equipped with adaptive modulation (AM). To maintain high image quality during transmission into the worst channel conditions, a bit error rate (BER) of 10-5 is chosen as a threshold. Figure 2 illustrates the theoretical performance of QPSK, 8PSK, 14QAM and 64QAM modulation schemes.

2C5-3

Figure 2. Theoretical Performance of Different Modulations

Intersection points between the curves of modulations performance and 10-5 BER horizontal line refers to the key switching point from a modulation to another. Table 1 shows the different switching thresholds for adaptive modulation.

Table 1. Switching Thresholds for Adaptive Modulation

Suitable Modulation Eb/N0 Range (dB)

64QAM Eb/N0 > 17.8

16QAM 13.4 < Eb/N0 ≤ 17.8

8PSK 12.8 < Eb/N0 ≤ 13.4

QPSK 9.6 < Eb/N0 ≤ 12.8

Principle of Operation Figure 3 illustrates the transmission operation

equipped by adaptive modulation approach. After converting the video captured to an appropriate GNU Radio data file format, data passes through an encoder to prepare the packets.

Figure 3. Block Diagram of Data Transmission with Adaptive Modulation

These packets are directly injected in the modulator in order to generate the analog signals corresponding to the selected modulation. These analog signals are transmitted via an antenna at a given center frequency f1 to the receiver (TX process).

At the reception part, the received signals are passed through several synchronization steps (time, phase and frequency) for good signal recovery following by filtration operations.

Thereafter, the signal is demodulated to rebuild packets. Finally, the packets go through a decoder to obtain the final received file (RX process).

Meanwhile, a real time signal to noise ratio (SNR) estimation is continuously carried out at the receiver. The (SNR) value provides information about current channel state. Hence, the demodulator is directly informed of the suitable modulation to use, whereas the modulator is informed to change the modulation through a feed-back control message sent by receiver.

Therefore, the modulator and the demodulator units can easily switch to an adequate modulation regardless of transmission conditions thanks to SNR estimation, enabling automatic adjustment of the system.

2C5-4

Adaptive Modulation Algorithm Switching thresholds for adaptive modulation

are given in Table 1 by the ratio of the energy per bit to the noise spectral density (Eb/N0). However, the channel state estimation is given by the SNR.

Therefore, the adaptive modulation algorithm can be written by converting the Eb/N0 value to SNR by using the relationship:

SNR = Eb/N0 + 10×log10 (number of bits per symbol)

The adaptive modulation algorithm is presented in Figure 4.

Figure 4. Adaptive Modulation Conceptual Diagram

IV. Test and Performance

Test Platform In order to test the adaptive modulation system

performance, the laboratory test bench contains two parts as shown in Figure 5:

• Transmitter, Tx: It consists of a ZeptoSDR platform connected to a laptop with Ethernet gigabit cable. Fitted with GNU radio software, the laptop is able to control transmission power, to display both the transmitted signals constellation and to feed-back the modulation control message. The adaptive modulation transmitter flowgraph is detailed in Figure 6.

• Receiver, Rx: It consists of a ZeptoSDR platform connected to a second laptop. However, this part is responsible for SNR estimation and receiving data. Therefore, the receiver flowgraph, given in Figure 7, allows displaying the received signal constellation, SNR estimated value and output data signal.

Note that the feed-back modulation control message is sent to the transmitter by using a Gaussian Minimum Shift Keying modulation (GMSK). This modulation is chosen thanks to its robustness, its phase shift keying continuity and its implementation simplicity.

2C5-5

Figure 5. Lab Test Bench Platform

Figure 6. Adaptive Modulation Transmitter

2C5-6

Figure 7. Adaptive Modulation Receiver

System Performance Results

Before testing the adaptive modulation transmission in L-Band frequency, every modulation scheme used was tested and validated. As shown in Figure 8, which illustrates the constellation diagrams for the different modulations implemented, the channel bandwidth can be efficiently employed by switching to higher modulation order under good channel conditions. This fact allows transmitting more data while maintaining the desired BER.

Figure 8. Adaptive Modulation Constellation Diagrams for QPSK, 8PSK, 16QAM and 64QAM

2C5-7

As presented in previous section, the SNR estimation is performed at the receiver. Hence, the demodulator is directly informed by the estimator in order to change the actual modulation used toward a suitable one. Figure 9 exhibits an instantaneous switching from QPSK to 8PSK modulation.

Figure 9. QPSK to 8PSK Switching at the Receiver

However, the transmitter receives the order of modulation change through a feed-back message. This fact causes a switching time delay between the transmitter and the receiver, as shown in Figure 10, which can lead to lose some packets. The delay value can be obtained based on the transmission speed and the amount of information lost. Experimental tests indicate that time delay is almost equal to 1.8s.

(a)

(b)

Figure 10. Switching Delay

The switching delay can be compensated by implementing a modulation time offset estimated at the receiver. The time offset allows the receiver to conserve the modulation used even receiving the switching order until the transmitter receives the switching information.

This solution is valid when the system passes from less efficient modulation to another modulation more efficient (Figure 10.a) thanks to suitable channel conditions for both modulations.

On the other hand, using time offset during switching modulation from an efficient modulation to a less efficient one (Figure 10.b) will cause errors (the higher efficient modulation will be used below its SNR threshold). To keep reliable transmission in this case, increasing the SNR thresholds for all modulations can be a solution. For instance, the system switches to 8PSK modulation when Eb/N0 reaches 13.4dB (the 16QAM threshold for reliable transmission).

If this threshold value is set as 13.6dB, the system will be able to keep reliability under 16QAM over the switching delay.

Finally, Figure 11 compares the BER theoretical performance and experimental one for a transmission equipped by adaptive modulation. Results are obtained by sending a static reference video and comparing the received file with the reference one.

Figure 11. Comparison of Theoretical and Experimental BER Performance

Theoretical and experimental curves show similar behavior. Slightly more errors can be observed on the experimental curves. The difference

2C5-8

can be explained by the fading phenomenon due to the multipath caused by movements in the laboratory experimentation.

V. Conclusion An experimental implementation of adaptive

modulation technique for air-to-ground (A/G) communication systems using a software defined radio approach was presented.

The system implementation provides a transmission, of a real time captured video, equipped by an adaptive modulation technique. The system autonomously chooses the most suitable modulation among BPSK, QPSK, 8PSK, 16QAM based on the estimation of link quality measured by the signal to noise ratio (SNR) in real time.

More investigations are currently done to improve 64QAM transmission quality in order to offer more bits per symbol.

In addition, adaptive coding is currently being implemented to have adaptive modulation and coding (AMC) for (A/G) communication systems.

References [1] L. Wan, H. Zhou, X. Xu, Y. Huang, S. Zhou, Z. Shi and J.H. Cui, April 2015, "Adaptive Modulation and Coding for Underwater Acoustic OFDM," in IEEE Journal of Oceanic Engineering, vol.40, no.2, pp.327-336.

[2] P.V.R. Ferreira, R. Metha and A.M. Wyglinski, Dec. 2014, "Cognitive Radio-Based Geostationary Satellite Communications for Ka-Band Transmissions," in 2014 IEEE Global Conference on Signal and Information Processing (GlobalSIP), pp.1093-1097.

[3] H. Meric, J. Lacan, F. Arnal, G. Lesthievent and M.L. Boucheret, Dec. 2013, "Combining Adaptive Coding and Modulation With Hierarchical Modulation in SatCom Systems," in IEEE Transactions on Broadcasting, vol.59, no.4, pp.627-637.

[4] D. Tarchi, G.E. Corazza and A. Vanelli-Coralli, June 2013, "Adaptive Coding and Modulation Techniques for Next Generation Hand-Held Mobile

Satellite Communications," in 2013 IEEE International Conference on Communications (ICC), pp.4504-4508.

[5] L. Jie, Ba. Zhongming, Z. Chuanxiong and L. Qiyue, April 2015, "Scalable Video Multicast with Joint Resource Allocation and Adaptive Modulation and Coding Over Multiple Base Station Networks," in IEEE 12th International Conference on Networking, Sensing and Control (ICNSC), pp.134-139.

[6] J. Yang, Y. Ran, S. Chen, W. Li and L. Hanzo, January 2015, "Online Source Rate Control for Adaptive Video Streaming Over HSPA and LTE-Style Variable Bitrate Downlink Channels," in IEEE Transactions on Vehicular Technology, pp.1,1

[7] Y. Zheng, S. Ren, X. Xu, Y. Si, M. Dong and J. Wu, March 2012, "A Modified ARIMA Model for CQI Prediction in LTE-Based Mobile Satellite Communications," in 2012 International Conference on Information Science and Technology (ICIST), pp.822-826.

[8] M. S. Ben Mahmoud, C. Guerber, N. Larrieu, A. Pirovano and J. Radzik, "Aeronautical Air−Ground Data Link Communications", John Wiley & Sons 2014.

[9] M. Raja, A.P. Vinod and A.S. Madhukumar, April 2015, "DME Interference Mitigation for LDACS1 Based on Decision-Directed Noise Estimation," in 2015 Integrated Communication, Navigation, and Surveillance Conference (ICNS), pp.21-23.

[10] Z. Chao, Z. Yaxin, X. Jialuo and J. Yu, May 2015, "Aeronautical Central Cognitive Broadband Air-to-Ground Communications," in IEEE Journal on Selected Areas in Communications, vol.33, no.5, pp.946-957.

[11] D. S. Ponchak, R. D. Apaza, B. Haynes, J. M. Wichgers and A. Roy, March 2015, "A Study of Future Communications Concepts and Technologies for the National Airspace System part IV", in 2015 IEEE Aerospace Conference, pp.7-14.

[12] Nutaq, “ZeptoSDR,”. Available: http://nutaq.com/en/products/zeptosdr.

2C5-9

[13] GNU Radio – Wiki Start [Online]. Available: http://www.gnuradio.org

[14] T. Jalloul, W. Ajib, O.A. Yeste-ojeda, R. Landry, C. Thibeault, Oct. 2014, "DME/DME Navigation Using a Single Low-Cost SDR and Sequential Operation," in 2014 IEEE/AIAA 33rd Digital Avionics Systems Conference (DASC), pp.5-9.

Acknowledgements This research project entitled “AVIO-505” was

supported by the Consortium for Research and Innovation in Aerospace in Quebec (CRIAQ), and with the collaboration of Bombardier Aerospace, MDA, Marinvent Corporation and Nutaq.

Email Addresses zakaria.el-alaoui-ismaili@ens.etsmtl.ca

ajib.wessam@uqam.ca

omar.yeste@lassena.etsmtl.ca

renejr.landry@etsmtl.ca

34th Digital Avionics Systems Conference September 13-17, 2015

2C5-10