Key Techniques Research on UAV Data Link

Available online at www.sciencedirect.com

ScienceDirect Procedia Engineering 99 (2015) 1099 – 1107

“APISAT2014”, 2014 Asia-Pacific International Symposium on Aerospace Technology, APISAT2014

Key Techniques Research on UAV Data Link Jinxi Li*, Yongfei Ding, Zheng Fang China National Aeronautical Radio Electronics Research Institute, Shanghai, 200241, China

Abstract Considering the environments and requirements of the UAV data link, this paper presents an adaptive information transmission method. When transmitting the remote control command data, spectrum spread system is used to improve reliability of the data link. When transmitting the high rate information such as videos and images, a three-processing-channel optimum strategy is used to improve the reliability in different channel conditions. Channel 1 is based on Gardner timing recovery and MCMA equalizer, Channel 2 is based on spread spectrum code recovery and MCMA equalizer, Channel 3 is based on Gardner timing recovery and carrier recovery. The strategy computes the confidence level of each process channel and chooses the best one as the output. Besides, the typical channel which the UAV encountered is analyzed and modeled. Experiments show that this design can achieve good performance in complex environments. © Published by Elsevier Ltd. This © 2015 2014The TheAuthors. Authors. Published by Elsevier Ltd.is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of Chinese Society of Aeronautics and Astronautics (CSAA). Peer-review under responsibility of Chinese Society of Aeronautics and Astronautics (CSAA)

Keywords: data link; Unmanned Aerial Vehicle (UAV); complex environments

1. Introduction The data link is an important part of the UAV, and with ground control station, it can complete the remote control of UAV. Meanwhile, the data link can transmit the airborne information such as videos and images. Generally, reflection, scattering, diffraction, and shadowing effects together with a direct line-of-sight (LOS) path are known as multipath propagation in communications and result in fading of the received signal due to constructive and deconstructive superposition. Therefore, it is necessary to establish an appropriate model for the aeronautical channel to analyze its performance. Firstly, this paper describes the channel characteristics of UAV communication,

* Corresponding author. Tel.: +86-21-33297977. E-mail address: [email protected]

1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of Chinese Society of Aeronautics and Astronautics (CSAA)

doi:10.1016/j.proeng.2014.12.645

1100

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

which is based on statistical models. And then an adaptive information rate method is presented. At last, simulations of key techniques are described. Test results show that this method can meet the requirement of the UAV data link, and the data link can work farthest stably in different environments. 2. UAV communication channel models[1,2] The propagation paths are composed of direct component r1, ground reflection components r2 and other reflection components r3, Fig.1 shows the channel model. In this paper, three-path model WSSUS (wide-sense stationary uncorrelated scattering) is analyzed in UAV air-ground data link [4, 5, 6]. A LOS channel with both specular and diffuse multipath is characterized by the impulse response:

h  t ,W a0e j 2S f LOS tG W  a1e

jTref

e

j 2S f ref t

G W  W ref  a2e jT e j 2S f tG W  W scat scat

scat

(1)

Where, a0, a1, a2respectivelypresentsthe amplitude gain of direct components, ground reflection components and other reflection components. fLOS is the Doppler shift component. Tref, fref, τref respectively presents phase shift, Doppler frequency shift and delay of ground reflection components.Tscat, fscat, τscat respectively presents phase shift, Doppler frequency shift and delay of other reflection components. In the glide state, there are LOS componentsr1and a strong multipath componentsr2 and r3, which are caused by surrounding buildings, mountains and other landforms around the airport. The fading events can be approximated by the Rician channel. The proposed Rice factor (K) is 6.9dB [3, 4]. The rural channel model of COST-207 is used in this state, the maximum multipath delay is 0.7μs, and the correlation bandwidth of channels is about 1.4MHz. If the information rate is above 2Mbps, the signal bandwidth is greater than the correlation bandwidth. The channel shows frequency selective fading.

r1

r2

r3 h2

mountain

h1

d

ground

Fig. 1 Channel model of A UAV air-ground information transmission

In the landing or takeoff state, the LOS signal r1is strengthened, r2 and r3is relatively weakened, and the main effect of multipath fading is caused by r2 and r3, wherein the delay of r2 is relatively small. For example, the antenna height of ground stations is 5m, the aircraft height is 0 to 5Km, the flight distance is 1 to 100Km, and the maximum multipath delay is about 16.7ns. In some flight environment, the multipath delay can be above 10μs. The proposed value of K is 15dB [4], and the correlation bandwidth is approximately 1MHz with maximum multipath delay of 1μs. If the information rate is above 2Mbps, the channel shows frequency selective fading too. In the cruise state, r3 is weaker than r1 and r2, and the energy of r3, which is about 1% to 8.4% of r1, can be ignored. In this state, the channel consists of r1 and r2. And the effect of multipath signals is mainly caused by r2. The K can be set to 10dB [4]. Assuming an ideal ground mirror reflection, the ground station antenna height is 5m, the aircraft height is 5 to 20Km, the flight distance is 100 to 200Km, and the multipath delay is about 0.42 to 6.67ns, and the minimum correlation bandwidth is approximately up to 150MHz. If the information rate is 2 to 64Mbps, the channel shows flat fading.

1101

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

3. Key technical analysis The design of the UAV data link requires an assessment of its performance and reliability in the desired operational environments. This paper focuses on the different environmental conditions that occur during the different flight states, and an adaptive transmission rate method is used to strengthen the system robustness. The key techniques include: adaptive information rate design, timing recovery, channel equalizer and so on. The main process of transmitter and receiver are shown separately in Fig. 2 and Fig. 3.

Fig. 2 Transmitter of UAV data link system ,XKW[KTI_ ULLYKZ JOYIXOSOTGZOTM

6-'

'*

**)

**)

3GZIN LORZKX

:OSOTMXKIU\KX_! INGTTKRKW[GRO`KX! ZNXKKINGTTKR VXKLKXXKJYZXGZKM_

+TIUJOTMGTJ JKLXGSOTM

U[ZV[Z

6U]KX KYZOSGZOTM '-)

Fig. 3 Receiver of UAV data link system

3.1. Adaptive information rate design The maximum RF bandwidth is determined in some UAV data link to satisfy the maximum information transmission rate. There may be differences due to different users, so the rate adaptation must be considered to send the information by using a unified communication frame. Fig.4 shows the structure of communication frame. HOZ ,XGSKNKGJKX

HOZ )'@')

HOZ *OYZGTIK 3KGY[XKSKTZ

HOZ KSVZ_VGIQKZ OJKTZOLOIGZOUT

HOZ H[YOTKYYJGZG

Fig. 4 communication frame structure

HOZ KSVZ_VGIQKZ OJKTZOLOIGZOUT

HOZ H[YOTKYYJGZG

1102

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

Where, the frame header is composed of the pseudo-random sequence, and this sequence is applied to frequency discrimination (coarse carrier synchronization), timing synchronization and frame synchronization. The complex CAZAC sequence is used for phase ambiguity correction, the next 240bit segment is used for distance measurement, the empty packet identification 1 and 2 is used to identify whether the business data segment 1 and 2 is valid. The business data segment1 and 2 is used for transmitting the encoded data with variable information in different channel conditions. When transmitting the aircraft status, uplink control commands, the information is important for the aircraft's control. A low bit rate data transmission method is used (12.8Kbps, 25.6Kbps, etc). And the data information is encoded by pseudo-random sequences combining with cascade correction coding. When transmitting videos and images, a high bit rate data transmission method is used (4.096Mbps, 8.192Mbps, etc).In this condition, the data information is the output of TPC coding. 3.2. Timing Recovery based on interpolation filter Fig.5 presents the timing recovery, which is mainly composed of interpolator [7, 8], timing error detector [9], loop filter [10] and number-controlled oscillator (NCO). The output of the interpolator is the product of the signal x(m) and the impulse response h(t), both of which have I adjacent samples. And the controller of Fig.5 is responsible for determining parameters of mk and μk , which is available to the interpolator. The necessary control is provided by the NCO, and the signal samples transferred through a shift register are synchronized to the output of NCO at a rate of l/Ts. Interpolator

mk

uk Controller

NCO

Timing error detector

Loop filter

Fig. 5 Timing Recovery Block Diagram

3.3. Channel Equalizer Generally, the receiver uses the channel equalizer to overcome the signal distortion caused by inter-symbol interference (ISI) [11, 12]. The modified constant modulus algorithm (MCMA) obtains the cost function of unsynchronized signals, and updates weight coefficients of the filter. The cost function is composed of signal’s modulus and phase information, so carrier recovery can be completed while the equalizer is working. When there is no frequency offset, MCMA can be corrected to recover the arbitrary phase rotation caused by channel fading. However, to some extent, MCMA can track the frequency offset in the presence of carrier frequency offset, and the weights of the equalizer are adjusted according to the channel characteristics. Fig.6 shows the constellation after timing recovery and after equalization, where, the flight speed is 700 Km/h and the multipath delay is 0.5μs. As shown in Fig. 6, the constellation is confused, and there are some multipath components after timing recovery, and the constellation is concentrated after equalization.

1103

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

Gardner timing recovery

Gardner timing recovery and MCMA equalizer

2.5

1.5

2 1

1.5 1 Quadrature

Quadrature

0.5 0.5 0 -0.5

0

-0.5 -1 -1.5

-1

-2 -2.5

-1.5 -2

-1

0 In-Phase

1

2

-1.5

-1

-0.5

0 In-Phase

0.5

1

1.5

Fig. 6 Signal constellation after timing recovery and after equalization

3.4. BER Performance of M-ary DSSS In this section, we analyzed the bit error rate (BER) performance of M-ary (M=16) direct sequence spread spectrum system (DSSS). Fig. 7 shows the BER performances tested in theoretical condition and real condition. The performance in the real condition has about 0.55dB demodulation loss at 10-5BER value. The loss is mainly caused by truncation effects of filters, finite word length effects of digital processing systems, timing and carrier frequency offset. BER of DSSS

0

10

Tested BER of 16-ary DSSS Theoretical BER of 16-ary DSSS Theoretical BER of 128 DSSS

-1

10

-2

BER

10

-3

10

-4

10

-5

10

-6

10

-19

-18

-17

-16

-15

-14 -13 Ec/N0(dB)

-12

-11

Fig. 7 BER of spread spectrum system

-10

-9

-8

1104

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

BER of spread spectrum system in Rician channels

-1

10

speed is 700km/h speed is 100km/h

-2

10

-3

10

-4

BER

10

-5

10

-6

10

-7

10

-8

10

-9

10

0

2

4

6

8 K factor/dB

10

12

14

16

Fig. 8 BER of spread spectrum system in Rician channels

Fig. 8 shows the BER performance of 16-ary DSSS in Rician channels. Where, the noise floor is 0dB, the flight speed is 100Km/h or 700Km/h. In the glide state, the flight speed is less than 100Km/h; the BER of the demodulation is about 10-4 at 6.9dB K value. After encoding, the BER performance can be improved to 10-5. In the landing and takeoff state, the flight speed is less than 100Km/h, the BER performance is better than 10-8 at 15dB K value. In the cruise state, the flight speed is about700Km/h, the BER performance is about 10-6 at 10dB K value. As shown in Fig.9, the Doppler shift caused by the flight speed has little effect on M-ary DSSS. Fig. 9 shows the simulation results separately in frequency selective fading channel and flat fading channel. The demodulation loss is higher in flat fading channel than that in frequency selective fading channel, because the signal detection is more difficult in frequency selective fading channel than that in flat fading channel. 0

Performance of 16-ary DSSS in frequency selective fading channels

10

Path loss:0dB Path loss:3dB Path loss:7.5dB Path loss:12dB AWGN

-1

10

-2

BER

10

-3

10

-4

10

-5

10

-6

10

-18

-16

-14

-12 Ec/N0(dB)

-10

-8

-6

1105

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

Performance of 16-ary DSSS in flat fading channels

0

10

Path loss:0dB Path loss:3dB Path loss:7.5dB Path loss:12dB AWGN

-1

10

-2

BER

10

-3

10

-4

10

-5

10

-6

10

-18

-16

-14

-12

-10 Ec/N0(dB)

-8

-6

-4

-2

Fig. 9 Performance of 16-ary DSSS in frequency selective fading channels and in flat fading channels

3.5. BER Performance of the three-processing-channel optimum strategy When transmitting high rate information, there are three signal processing channels to be chosen. Channel 1 is based on Gardner timing recovery and MCMA equalizer, Channel 2 is based on spread spectrum code recovery and MCMA equalizer, Channel 3 is based on Gardner timing recovery and carrier recovery. The output is chosen based on the BER performance of the three processing channels. If the BER performance of each processing channel is within the same range, the channel with highest confidence level will be chosen. Fig.10 and Fig.11 show the results of using three-processing-channel optimum strategy in different channel conditions. Where, the multipath delay is supposed to be 0.5μs in frequency selective fading channel and to be 0.1μs in flat fading channel. The two figures show that the equalizer plays an important role in fading channels. Performance of high bit rate transmission in frequency selective fading channels 0 10 Channel 1 Channel 2 -1 Channel 3 10 Channel selected -2

BER

10

-3

10

-4

10

-5

10

-6

10

0

2

4

6 Path loss/dB

8

10

12

1106

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

Performance of high bit rate transmission in flat fading channels

0

10

Channel Channel Channel Channel

-1

10

1 2 3 selected

-2

BER

10

-3

10

-4

10

-5

10

-6

10

0

2

4

6 Path loss/dB

8

10

12

Fig. 10 Performance of high bit rate transmission in frequency selective fading channels and in flat fading channels Performance of high bit rate transmission in AWGN channels

-1

10

Channel 1 Channel 2 Channel selected Theory

-2

10

-3

BER

10

-4

10

-5

10

-6

10

0

2

4

6 Eb/N0(dB)

8

10

12

Fig. 11 Performance of high bit rate transmission in AWGN channels

4. Conclusion Within the complex channel condition, UAV data link often encounters obstacle caused by multipath and Doppler, such as the worsening of CNR, fading channel, etc. Modern information warfare such as Net Centric Warfare requires the UAVs data link having good performance in different channel condition. The airborne military data link must be much more flexible and reliable. An adaptive information transmission method is mainly described, and the BER performance of this method is discussed in different channel models. The simulation and test results show that the proposed method has good performance in different channel conditions such as AWGN and fading channels. The three-processing-channel optimum strategy can improve the reliability of UAV data link in complex environments, and this paper presents some technical basis for applications of airborne military data link.

Jinxi Li et al. / Procedia Engineering 99 (2015) 1099 – 1107

References [1] Gao Baosheng, Channel Characteristics of UAV Ground to air Data link and Broadband Data Transmission, Radio Engineering, Vo1.41, No.9, Sep. 2011, pp.4-6. [2] Jin Shi, A Statistical Model for the UAV Communication Channel, Acta Aeronautica Et Astronautica Sinica, Vol.25, No.1, Ja n.2004,pp. 6265. [3] Hoeher P, A statistical discrete time model for the WSSUS multipath channel, IEEE Trans Veh Techmol, Vol.41, No.4, 1992, pp. 461-468. [4] Zheng Jinxiu, Fading Characteristics of UAV TT&C Channel and Anti-fading Technology, Telecommunication Engineering, Vol.51,No.3, Mar.2011, pp. 56-58. [5] Michael RICE, Arrow band Channel Model for Aeronautical for Aeronautical Telemetry, IEEE Trans on AEROSPACE AND ELECTRONIC SYSTEMS, Vol.36, No.4, 2000, pp. 1371-1377. [6] Michael RICE, Wideband channel model for aeronautical telemetry, IEEE Transactions on aerospace and Electronic systems, Vol.40, No.1, 2004, pp. 57~69. [7] GARDNER F M, A BPSK/QPSK timing-error detector for sampled receivers, IEEE Trans Communication, Vol.46, No.5, 1986, pp. 423-429. [8] GARDNER F M, Interpolation in digital modems part I: fundamentals, IEEE Trans Communication, Vol.41, No.3, 1993, pp. 501-507. [9] ERUP L, GARDNER F M, HARRIS R A, Interpolation in digital modems part II: implementation and performance, IEEE Trans Communication, Vol.41, No.5, 1993, pp. 998-1008. [10] Fu Yongming, Parameters design and performance analysis of the timing recovery loop based on Gardner timing detector, Jo urnal on Communications, Vol.33, No. 6, 2012, pp. 191-198. [11] SHAHID UH QURESHI, Adaptive Equalization, IEEE PROCEEDINGS, Vol.73, No.9, 1985, pp. 1349-1387. [12] Mahmood Farhan Mosleh, Combination of LMS and RLS Adaptive Equalizer for Selective Fading Channel, European Journal of Scientific Research, Vol.43, No.1, 2010, pp. 127-137.

1107