- Email: [email protected]
Experimental demonstration of single carrier underwater acoustic communication using a vector sensor
Applied Acoustics 98 (2015) 1–5
Contents lists available at ScienceDirect
Applied Acoustics journal homepage: www.elsevier.com/locate/apacoust
Experimental demonstration of single carrier underwater acoustic communication using a vector sensor Xiao Han, Jing-wei Yin ⇑, Ge Yu, Peng-yu Du Acoustic Science and Technology Laboratory, Harbin Engineering University, Harbin 150001, China College of Underwater Acoustic Engineering, Harbin Engineering University, Harbin 150001, China
a r t i c l e
i n f o
Article history: Received 27 December 2014 Received in revised form 17 March 2015 Accepted 20 March 2015 Available online 16 May 2015 Keywords: UWA communications Single carrier Vector sensor Time reversal
a b s t r a c t Vector sensors could simultaneously collect not only pressure but also velocity signals and there will be a valuable gain when combining these two kinds of signals properly. This paper studies single carrier underwater acoustic (UWA) communication using a vector sensor. In a recent field experiment, acoustic communication transmissions were carried out over a 1 km range in shallow water at Songhua River, Heilongjiang province, China. The same coded signal was transmitted when the vector sensor was deployed to depth of 1 m, 1.5 m, 2 m, and 2.5 m respectively. Decoding results, that nearly error-free performance at four depths, show the robustness of UWA communications using a single vector sensor in a highly refractive environment. The four received signals are also combined together to asynchronously realize a vertical array processing using passive time reversal based on the spatial and temporal diversity, achieving a maximum gain about 8.02 dB compared with a single sensor processing. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Ocean sampling is now playing a more and more important role in the field of marine environment and global climate monitoring. While data from satellites and airplanes can provide with some useful information, it does not present information far below the ocean surface [1]. Variable undersea observation equipments such as autonomous underwater vehicles (AUVs) have emerged over the last decade to further understand the internal changes of ocean. Undersea observation equipments rely on high data rate UWA communications to transmit huge amount of data to the base station in real-time. However, the UWA channel is generally recognized as one of the most difficult communication channels today [2,3]. It has very severe delay spreads which could be on the order of 100 ms in shallow water and varies fast both in spatial and temporal domain. The useful bandwidth of UWA channel is very limited due to large transmission loss related to the carrier frequency. All these factors make it very difficult to communicate in UWA environment, especially for high data rate communications. Orthogonal Frequency Division Multiplexing (OFDM) has been a hot topic in high data rate UWA communications for its good
⇑ Corresponding author at: Acoustic Science and Technology Laboratory, Harbin Engineering University, Harbin 150001, China. Tel.: +86 13674664846. E-mail address: [email protected] (J.-w. Yin). http://dx.doi.org/10.1016/j.apacoust.2015.03.019 0003-682X/Ó 2015 Elsevier Ltd. All rights reserved.
performance in anti inter symbol interference (ISI) and bandwidth efficiency. Researchers all over the world have conducted a lot of simulations and experiments based on OFDM, yielding promising results [4–7]. However, there are also some known drawbacks. OFDM signal has higher peak-to-average power ratio (PAPR) which makes it difficult to transmit using transducers. As subcarriers are densely spaced in frequency domain, Doppler shifts due to relative motion between the source and the receiver will damage orthogonality of subcarriers, causing severe inter carrier interference (ICI). Single carrier communication outperforms OFDM in terms of signal PAPR and sensitivity to Doppler shifts. It has been widely used in UWA communications since Stojanovic et al. [8] introduced decision feedback equalizer (DFE) embedded with phase lock loop (PLL) into this scheme, solving the problem of phase tracking in 1994. The computation complexity of DFE in single carrier communications is very large as the taps should span the whole severe multipath spread. Frequency domain equalization [9–11] which adopts FFT and IFFT to realize channel equalization instead of DFE has been proposed in order to decrease the system complexity. In most cases, using only a single sensor is not enough to recover transmitted signals for long range communications. So a vertical array with multiple sensors is always used in single carrier communications. Song et al. [12–14] studied time reversal DFE which is proved to have a similar performance with multichannel DFE [15,16]. Both methods exploit spatial diversity achieved by a vertical array to get nearly ideal results.
2
X. Han et al. / Applied Acoustics 98 (2015) 1–5
The motivation of this paper is to investigate UWA communications using a single vector sensor. Vector sensors could collect not only pressure signals but also velocity signals and there will be a valuable gain when combining these two kinds of signals properly. To accomplish this objective, a field experiment (SHJ14) was conducted in Songhua river, Heilongjiang province, China in November 2014. Section 2 reviews the signal processing method using a vector sensor and time reversal, proposes the scheme in this paper. Section 3 describes the experiment environment. Section 4 presents decoding results using vector sensor. Section 5 is a brief summary.
intensity (AAAI) [17] is used in this paper to estimate source bearing. Fig. 2 shows the structure diagram of AAAI. Suppose the local reference signal is r(t), synchronous signals extracted in vx(t) and vy(t) will be matched filter with r(t), and then select the maximum value. The output of AAAI after peak selector
(
Ix ¼ max½rðtÞv x ðt sÞ ¼ max½C v x ;r ðsÞ ¼ As cos h þ Dx Iy ¼ max½rðtÞv y ðt sÞ ¼ max½C v y ;r ðsÞ ¼ As sin h þ Dy
where As is the correlation peak; Dx and Dy are small interference. So the estimated bearing is
^h ¼ arctan Ix Iy
2. Single carrier communication using a vector sensor
ð4Þ
ð5Þ
2.1. Signal model of vector sensor 3. SHJ14 experiment Vector sensor is made up of a traditional pressure sensor which has no directivity and three velocity sensors. It could collect not only pressure but also orthogonal velocity information in the underwater acoustic environment. So it has more inputs compared T with pressure sensors. Denote sðtÞ ¼ ½ pðtÞ v x ðtÞ v y ðtÞ v z ðtÞ as the output signals of vector sensor, where p(t) is the pressure signal and vx(t), vy(t), vz(t) are three velocity signals. Acoustic represents as standing wave in the vertical direction in ocean waveguide. So only p(t), vx(t), vy(t) can be used in underwater acoustic communications. Consider a two-dimensional condition, received signals of vector sensor can be described as
8 > < pðtÞ ¼ xðtÞ v x ðtÞ ¼ xðtÞ cos h > : v y ðtÞ ¼ xðtÞ sin h
ð1Þ
where h is the horizontal azimuth of acoustic arrivals and h e [0, 2p] Dipole directivity of vector sensors can be electronic rotated. Donate vc(t) and vs(t) as the combined directivity after electronic rotated, then
v c ðtÞ ¼ v x ðtÞ cos u þ v y ðtÞ sin u ¼ xðtÞ cosðh uÞ v s ðtÞ ¼ v x ðtÞ sin u þ v y ðtÞ cos u ¼ xðtÞ sinðh uÞ
ð2Þ
where u is the guide bearing. A linear combination (i.e. p + a vc) of p(t) and vc(t) is usually used in underwater acoustic communication aiming to get more processing gain. It is necessary to discuss the processing gain of p + a vc here. Suppose signal power and noise power in p(t) are r2s and r2n respectively. Signal power and noise power in vc(t) will be r2s and r2n =2. So the SNR of p + a vc can be described as
ðS=NÞpv ¼
ð1 þ aÞ2 r2s 1 þ a2 =2 r2n
ð3Þ
The SNR will reach a maximum value when a = 2. So in the following processing of experimental data p + 2 vc will be deployed. 2.2. Single carrier communication scheme Fig. 1(a) depicts the single carrier communication scheme using a single vector sensor. It mainly consists of two parts: vector signal processing and DFE. Vector sensor processing could get input gain through combining p(t) and vc(t); DFE is used to compress multipath spread. Fig. 1(b) is the scheme when multiple sensors could be used. Input signals are firstly combined at every channel and then time reversal mirror (TRM) is used to compress the multipath, following a single channel DFE to remove residual ISI. A key part in Fig. 1 is the source bearing estimator. Active average acoustic
SHJ14 based on single carrier underwater acoustic communication [12] was conducted at Songhua river, Heilongjiang province, China in November 2014. The communication sequences are modulated using QPSK. The transmitted waveform is pulse shaped using a raised-cosine filter with a roll-off factor 0.7. The center frequency is 12 kHz, the symbol rate is 2 k symbols/s, and all signals are sampled at 96 kHz. 26,000 bits data were transmitted during this experiment and the first 1000 bits were used as training (NT = 500). The schematic of SHJ14 is shown in Fig. 3. The experiment involves one transducer, denoted by T whose source level was 186 dB re 1 lPa at 1 m, having a bandwidth of 8–20 kHz. The transducer was deployed to 1.5 m depth. A vector sensor, 1000 m away from the source, was deployed off the receiving boat which was always moored during this experiment. Note that the vector sensor was rigidly connected to the receiving boat through a steel pipe whose diameter was about 10 cm. The vector sensor was orderly deployed to point A, B, C, and D respectively. The same signal was transmitted at every depth and received signals were denoted by VS01, VS02, VS03, and VS04 in sequence. These collected signals will be used to demonstrate the decoding advantages of a single vector sensor processing. Fractionally spaced equalizers (two samples per symbol) were used for feed forward filters, and the number of feed forward (FF) and feedback (FB) filter taps were both 20. The proportional and integral phase tracking constants for the PLL were P1 = 0.0001 and P2 = 0.00001, respectively. The recursive least square (RLS) algorithm was employed to adaptively update the DFE equalizer coefficients with a forgetting factor k ¼ 0:9965. 4. Performance analysis In addition to transmitting the coded signal, linear frequency modulated (LFM) chirps, 50 ms in duration, and 10–14 kHz in frequency were also sent as sync signals. The chirp sequences were used to detect channel impulse response (CIR) from the source to vector sensor at different depths. The received signals are firstly match filtered by the local sync signal, that LFM chirps. Fig. 4 shows the output after match filtering the received signals VS01, VS02, VS03, and VS04 respectively. From Fig. 4 the following conclusions can be draw: (1) the multipath spread is less than 5 ms, not very severe at communication range 1000 m; (2) the multipath structure is not clear and acoustic arrivals from the source mixed together, indicating that the experimental environment is highly refractive. The angles estimated by AAAI method for VS01, VS02, VS03, and VS04 are 35°, 37°, 34°, and 33° respectively. The differences between estimated angles
X. Han et al. / Applied Acoustics 98 (2015) 1–5
3
Fig. 1. Single carrier communication scheme using vector sensor (a) a single vector sensor processing; (b) TRM processing using multiple vector sensors, time reversal combining is based on knowledge of the channel responses hi(t) and is followed by a single channel equalizer to remove the residual ISI.
Fig. 2. The structure diagram of AAAI synchronous signals in
Fig. 3. Schematic of the single carrier underwater acoustic communication experiment (SHJ14) conducted in Songhua River, Heilongjiang province, China in November 2014.
vx(t) and vy(t) are used to estimate the bearing of source.
are mainly due to the inaccurate movement of the steel pipe which may slightly change posture of the vector sensor. Decoding performance using a single vector sensor is shown in Fig. 5. Communications performance is quantified in terms of the output signal to noise ratio (SNR) and the bit error rate (BER) in the soft demodulation output. Fig. 5(a) is the decoding constellation for VS01; Fig. 5(b) is the decoding constellation for VS02; Fig. 5(c) is the decoding constellation for VS03; Fig. 5(d) is the decoding constellation for VS04. Table 1 shows the statistical experimental results. Though there was no sound speed profiler available during this field experiment, one could easily know the sound speed was a
4
X. Han et al. / Applied Acoustics 98 (2015) 1–5 Table 1 The experimental results.
Fig. 4. Channel impulse response detected from different received signals during the SHJ14 experiment at communication distance of 1000 m (a) VS01, the depth of vector sensor is 2.5 m; (b) VS02, the depth of vector sensor is 2.0 m; (a) VS03, the depth of vector sensor is 1.5 m; (a) VS04, the depth of vector sensor is 1.0 m.
slightly positive gradient distribution given that the environment temperature was around zeros degree Celsius. So there would be more acoustic energy in the area near the river surface. It can also be seen from Fig. 5 that the decoding results of VS01 and VS02 are much better than that of VS03 and VS04 which proves the above inference. Another interesting phenomenon is that decoding output SNRs of VS02 and VS03 are much higher than that of VS01 and VS04 respectively. An obvious reason is the difference between
Signals
Output SNR (dB)
BER (%)
VS01 VS02 VS03 VS04
9.90 12.94 9.25 8.75
0.21 (55/26,000) 0 (0/26,000) 0.41 (107/26,000) 0.62 (160/26,000)
their channels. As point C and point B located in the middle of the water column (depicted in Fig. 3), received signals at these two points suffered minor influence from the surface and bottom, so the multipath structures were much simpler. Considering the depth of experimental water column is only 6 m, the decoded results are very reasonable. Fig. 6 shows the decoding constellation using TRM when asynchronously combining the four received signals VS01, VS02, VS03, and VS04 together based on the spatial and temporal diversity. The output SNR = 16.77 dB and BER = 0(0/ 26,000). Several comments are in order. As observed in Figs. 5 and 6, the combination of P, vx, vy has more ability to combat the multipath of UWA channels, so a good performance is achieved even using a single vector sensor. There is a significant improvement in performance when decoding from using a single vector sensor to TRM which asynchronously combines the received signals together. The output SNR has been improved by 6.87 dB, 3.83 dB, 7.52 dB, and 8.02 dB for VS01, VS02, VS03, and VS04 respectively. These results show that multichannel processing could get valuable gains based on spatial and temporal diversity. Khosla and Dowling [18] pointed out that there would be a gain of 10 log M (M = 4 in this
Fig. 5. The decoding constellation of received signal VS01, VS02, VS03, and VS04 using a single vector sensor. (a) VS01: point A, the depth is 2.5 m; (b) VS02: point B, the depth is 2.0 m; (c) VS03: point C, the depth is 1.5 m; (d) VS04: point D, the depth is 1.0 m.
5
X. Han et al. / Applied Acoustics 98 (2015) 1–5
the China Postdoctoral 2013M531015).
Science
Foundation
(Grant
No.
References
Fig. 6. The decoding constellation using TRM when asynchronously combining received signal VS01, VS02, VS03, and VS04 together based on spatial and temporal diversity.
paper), 6.02 dB in theory when decoding using TRM with pressure sensors. The gains here are much higher than that as the combination of p(t) and v(t) could get additional gains. 5. Summary This paper presents a new underwater acoustic communication scheme, which utilizes vector sensor processing and time reversal technology. Single carrier underwater acoustic communication using a vector sensor is investigated based on experimental data collected in shallow water. The same signal was transmitted four times when the vector sensor was deployed to different depths. Decoding results show that even a single channel processing could get robust performance in a highly refractive environment. When asynchronously combining the received signals using TRM based on spatial and temporal diversity, additional 6.87 dB, 3.83 dB, 7.52 dB, and 8.02 dB gains are obtained for VS01, VS02, VS03, and VS04 respectively. Communications using vector sensor with Doppler shifts will be a further research topic in the future. Acknowledgements This research is supported by the National Natural Science Foundation of China (Grant Nos. 51179034, 61471137), Ships Pre-research Support Technology Fund (Grant No. 13J3.1.5), and
[1] Song HC, Howe Bruce M, Brown Michael G, Andrew Rex K. Diversity-based acoustic communication with a glider in deep water. J Acoust Soc Am 2014;135(3):EL1023–6. [2] Ying Xiao, Fu-liang Yin. Blind equalization based on RLS algorithm using adaptive forgetting factor for underwater acoustic channel. China Ocean Eng 2014;28(3):401–8. [3] Huang SH, Tsao J, Yang TC, Cheng SW. Model-based signal subspace channel tracking for correlated underwater acoustic communication channels. IEEE J. Ocean Eng. 2014;39(2):343–56. [4] Leus Geert, van Walree Paul A. Multiband OFDM for covert acoustic communications. IEEE J Sel Areas Commun 2008;26(9):1662–73. [5] Tao Jun, Zheng Yahong Rosa, Xiao Chengshan, Yang Robust TC. MIMO underwater acoustic communications using turbo block decision-feedback equalization. IEEE J Ocean Eng 2010;35(4):948–60. [6] Huang Jianzhong, Zhou Shengli, Huang Jie. Progressive inter-carrier interference equalization for OFDM transmission over time-varying underwater acoustic channels. IEEE J Sel Topic Signal Process 2011;5(8):1524–36. [7] Polprasert Chantri, Ritcey James A, Stojanovic Milica. Capacity of OFDM systems over fading underwater acoustic channels. J Ocean Eng 2011;36(4):514–24. [8] Stojanovic Milica, Catipovic Josko A, Proakis John G. Phase-coherent digital communications for underwater acoustic channels. IEEE J Ocean Eng 1994;19(1):100–11. [9] Xin Zhang, Xiao-Ji Zhang, Xiao-Fei Xing, Li-Wei Jiang. Frequency domain response estimation of underwater acoustic channel and noise estimation for single-carrier frequency domain equalization. Acta Phys Sin 2014;63(19). 194304_1-194304_8. [10] Xia Menglu, Rouseff Daniel, Ritcey James A, Zou Xiang, Chantri Polprasert, Xu Wen. Underwater acoustic communication in a highly refractive environment using SC–FDE 2014;39(3):491–9. [11] Cheng-Bing He, Jian-Guo Huang, Qing-Wei Meng, Qun-Fei Zhang, Wen-Tao Shi. PN-based single carrier block transmission with iterative frequency domain equalization over underwater acoustic channels. Acta Phys Sin 2013;62(23). 234301_1–234301_7. [12] Song HC, Hodgkiss WS. Efficient use of bandwidth for underwater acoustic communication. J Acoust Soc Am 2013;134(2):EL905–8. [13] Shimura T, Ochi H, Song HC. Experimental demonstration of multiuser communication in deep water using time reversal. J Acoust Soc Am 2013;134(4):3223–9. [14] Song HC. Equivalence of adaptive time reversal and least squares for cross talk mitigation. J Acoust Soc Am 2014;135(3):EL154–8. [15] Stojanovic M, Proakis JG, Catipovic JA. Performance of high-rate adaptive equalization on a shallow water acoustic channel. J Acoust Soc Am 1996;100(4):2213–9. [16] Flynn John A, Ritcey James A, Rouseff Daniel, Fox Warren LJ. Multichannel equalization by decision-directed passive phase conjugation: experimental results. IEEE J Ocean Eng 2004;29(3):824–36. [17] Jingwei Yin, Sen Yang, Pengyu Du, Yun Yu, Yang Chen. Code divided multiple access underwater acoustic communication based on active average acoustic intensity. Acta Phys Sin 2012;61(6). 064302_1–064302_7. [18] Khosla Sunny R, Dowling David R. Time-reversing array retrofocusing in noisy environments. J Acoust Soc Am 2001;109(2):538–46.
Contents lists available at ScienceDirect
Applied Acoustics journal homepage: www.elsevier.com/locate/apacoust
Experimental demonstration of single carrier underwater acoustic communication using a vector sensor Xiao Han, Jing-wei Yin ⇑, Ge Yu, Peng-yu Du Acoustic Science and Technology Laboratory, Harbin Engineering University, Harbin 150001, China College of Underwater Acoustic Engineering, Harbin Engineering University, Harbin 150001, China
a r t i c l e
i n f o
Article history: Received 27 December 2014 Received in revised form 17 March 2015 Accepted 20 March 2015 Available online 16 May 2015 Keywords: UWA communications Single carrier Vector sensor Time reversal
a b s t r a c t Vector sensors could simultaneously collect not only pressure but also velocity signals and there will be a valuable gain when combining these two kinds of signals properly. This paper studies single carrier underwater acoustic (UWA) communication using a vector sensor. In a recent field experiment, acoustic communication transmissions were carried out over a 1 km range in shallow water at Songhua River, Heilongjiang province, China. The same coded signal was transmitted when the vector sensor was deployed to depth of 1 m, 1.5 m, 2 m, and 2.5 m respectively. Decoding results, that nearly error-free performance at four depths, show the robustness of UWA communications using a single vector sensor in a highly refractive environment. The four received signals are also combined together to asynchronously realize a vertical array processing using passive time reversal based on the spatial and temporal diversity, achieving a maximum gain about 8.02 dB compared with a single sensor processing. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Ocean sampling is now playing a more and more important role in the field of marine environment and global climate monitoring. While data from satellites and airplanes can provide with some useful information, it does not present information far below the ocean surface [1]. Variable undersea observation equipments such as autonomous underwater vehicles (AUVs) have emerged over the last decade to further understand the internal changes of ocean. Undersea observation equipments rely on high data rate UWA communications to transmit huge amount of data to the base station in real-time. However, the UWA channel is generally recognized as one of the most difficult communication channels today [2,3]. It has very severe delay spreads which could be on the order of 100 ms in shallow water and varies fast both in spatial and temporal domain. The useful bandwidth of UWA channel is very limited due to large transmission loss related to the carrier frequency. All these factors make it very difficult to communicate in UWA environment, especially for high data rate communications. Orthogonal Frequency Division Multiplexing (OFDM) has been a hot topic in high data rate UWA communications for its good
⇑ Corresponding author at: Acoustic Science and Technology Laboratory, Harbin Engineering University, Harbin 150001, China. Tel.: +86 13674664846. E-mail address: [email protected] (J.-w. Yin). http://dx.doi.org/10.1016/j.apacoust.2015.03.019 0003-682X/Ó 2015 Elsevier Ltd. All rights reserved.
performance in anti inter symbol interference (ISI) and bandwidth efficiency. Researchers all over the world have conducted a lot of simulations and experiments based on OFDM, yielding promising results [4–7]. However, there are also some known drawbacks. OFDM signal has higher peak-to-average power ratio (PAPR) which makes it difficult to transmit using transducers. As subcarriers are densely spaced in frequency domain, Doppler shifts due to relative motion between the source and the receiver will damage orthogonality of subcarriers, causing severe inter carrier interference (ICI). Single carrier communication outperforms OFDM in terms of signal PAPR and sensitivity to Doppler shifts. It has been widely used in UWA communications since Stojanovic et al. [8] introduced decision feedback equalizer (DFE) embedded with phase lock loop (PLL) into this scheme, solving the problem of phase tracking in 1994. The computation complexity of DFE in single carrier communications is very large as the taps should span the whole severe multipath spread. Frequency domain equalization [9–11] which adopts FFT and IFFT to realize channel equalization instead of DFE has been proposed in order to decrease the system complexity. In most cases, using only a single sensor is not enough to recover transmitted signals for long range communications. So a vertical array with multiple sensors is always used in single carrier communications. Song et al. [12–14] studied time reversal DFE which is proved to have a similar performance with multichannel DFE [15,16]. Both methods exploit spatial diversity achieved by a vertical array to get nearly ideal results.
2
X. Han et al. / Applied Acoustics 98 (2015) 1–5
The motivation of this paper is to investigate UWA communications using a single vector sensor. Vector sensors could collect not only pressure signals but also velocity signals and there will be a valuable gain when combining these two kinds of signals properly. To accomplish this objective, a field experiment (SHJ14) was conducted in Songhua river, Heilongjiang province, China in November 2014. Section 2 reviews the signal processing method using a vector sensor and time reversal, proposes the scheme in this paper. Section 3 describes the experiment environment. Section 4 presents decoding results using vector sensor. Section 5 is a brief summary.
intensity (AAAI) [17] is used in this paper to estimate source bearing. Fig. 2 shows the structure diagram of AAAI. Suppose the local reference signal is r(t), synchronous signals extracted in vx(t) and vy(t) will be matched filter with r(t), and then select the maximum value. The output of AAAI after peak selector
(
Ix ¼ max½rðtÞv x ðt sÞ ¼ max½C v x ;r ðsÞ ¼ As cos h þ Dx Iy ¼ max½rðtÞv y ðt sÞ ¼ max½C v y ;r ðsÞ ¼ As sin h þ Dy
where As is the correlation peak; Dx and Dy are small interference. So the estimated bearing is
^h ¼ arctan Ix Iy
2. Single carrier communication using a vector sensor
ð4Þ
ð5Þ
2.1. Signal model of vector sensor 3. SHJ14 experiment Vector sensor is made up of a traditional pressure sensor which has no directivity and three velocity sensors. It could collect not only pressure but also orthogonal velocity information in the underwater acoustic environment. So it has more inputs compared T with pressure sensors. Denote sðtÞ ¼ ½ pðtÞ v x ðtÞ v y ðtÞ v z ðtÞ as the output signals of vector sensor, where p(t) is the pressure signal and vx(t), vy(t), vz(t) are three velocity signals. Acoustic represents as standing wave in the vertical direction in ocean waveguide. So only p(t), vx(t), vy(t) can be used in underwater acoustic communications. Consider a two-dimensional condition, received signals of vector sensor can be described as
8 > < pðtÞ ¼ xðtÞ v x ðtÞ ¼ xðtÞ cos h > : v y ðtÞ ¼ xðtÞ sin h
ð1Þ
where h is the horizontal azimuth of acoustic arrivals and h e [0, 2p] Dipole directivity of vector sensors can be electronic rotated. Donate vc(t) and vs(t) as the combined directivity after electronic rotated, then
v c ðtÞ ¼ v x ðtÞ cos u þ v y ðtÞ sin u ¼ xðtÞ cosðh uÞ v s ðtÞ ¼ v x ðtÞ sin u þ v y ðtÞ cos u ¼ xðtÞ sinðh uÞ
ð2Þ
where u is the guide bearing. A linear combination (i.e. p + a vc) of p(t) and vc(t) is usually used in underwater acoustic communication aiming to get more processing gain. It is necessary to discuss the processing gain of p + a vc here. Suppose signal power and noise power in p(t) are r2s and r2n respectively. Signal power and noise power in vc(t) will be r2s and r2n =2. So the SNR of p + a vc can be described as
ðS=NÞpv ¼
ð1 þ aÞ2 r2s 1 þ a2 =2 r2n
ð3Þ
The SNR will reach a maximum value when a = 2. So in the following processing of experimental data p + 2 vc will be deployed. 2.2. Single carrier communication scheme Fig. 1(a) depicts the single carrier communication scheme using a single vector sensor. It mainly consists of two parts: vector signal processing and DFE. Vector sensor processing could get input gain through combining p(t) and vc(t); DFE is used to compress multipath spread. Fig. 1(b) is the scheme when multiple sensors could be used. Input signals are firstly combined at every channel and then time reversal mirror (TRM) is used to compress the multipath, following a single channel DFE to remove residual ISI. A key part in Fig. 1 is the source bearing estimator. Active average acoustic
SHJ14 based on single carrier underwater acoustic communication [12] was conducted at Songhua river, Heilongjiang province, China in November 2014. The communication sequences are modulated using QPSK. The transmitted waveform is pulse shaped using a raised-cosine filter with a roll-off factor 0.7. The center frequency is 12 kHz, the symbol rate is 2 k symbols/s, and all signals are sampled at 96 kHz. 26,000 bits data were transmitted during this experiment and the first 1000 bits were used as training (NT = 500). The schematic of SHJ14 is shown in Fig. 3. The experiment involves one transducer, denoted by T whose source level was 186 dB re 1 lPa at 1 m, having a bandwidth of 8–20 kHz. The transducer was deployed to 1.5 m depth. A vector sensor, 1000 m away from the source, was deployed off the receiving boat which was always moored during this experiment. Note that the vector sensor was rigidly connected to the receiving boat through a steel pipe whose diameter was about 10 cm. The vector sensor was orderly deployed to point A, B, C, and D respectively. The same signal was transmitted at every depth and received signals were denoted by VS01, VS02, VS03, and VS04 in sequence. These collected signals will be used to demonstrate the decoding advantages of a single vector sensor processing. Fractionally spaced equalizers (two samples per symbol) were used for feed forward filters, and the number of feed forward (FF) and feedback (FB) filter taps were both 20. The proportional and integral phase tracking constants for the PLL were P1 = 0.0001 and P2 = 0.00001, respectively. The recursive least square (RLS) algorithm was employed to adaptively update the DFE equalizer coefficients with a forgetting factor k ¼ 0:9965. 4. Performance analysis In addition to transmitting the coded signal, linear frequency modulated (LFM) chirps, 50 ms in duration, and 10–14 kHz in frequency were also sent as sync signals. The chirp sequences were used to detect channel impulse response (CIR) from the source to vector sensor at different depths. The received signals are firstly match filtered by the local sync signal, that LFM chirps. Fig. 4 shows the output after match filtering the received signals VS01, VS02, VS03, and VS04 respectively. From Fig. 4 the following conclusions can be draw: (1) the multipath spread is less than 5 ms, not very severe at communication range 1000 m; (2) the multipath structure is not clear and acoustic arrivals from the source mixed together, indicating that the experimental environment is highly refractive. The angles estimated by AAAI method for VS01, VS02, VS03, and VS04 are 35°, 37°, 34°, and 33° respectively. The differences between estimated angles
X. Han et al. / Applied Acoustics 98 (2015) 1–5
3
Fig. 1. Single carrier communication scheme using vector sensor (a) a single vector sensor processing; (b) TRM processing using multiple vector sensors, time reversal combining is based on knowledge of the channel responses hi(t) and is followed by a single channel equalizer to remove the residual ISI.
Fig. 2. The structure diagram of AAAI synchronous signals in
Fig. 3. Schematic of the single carrier underwater acoustic communication experiment (SHJ14) conducted in Songhua River, Heilongjiang province, China in November 2014.
vx(t) and vy(t) are used to estimate the bearing of source.
are mainly due to the inaccurate movement of the steel pipe which may slightly change posture of the vector sensor. Decoding performance using a single vector sensor is shown in Fig. 5. Communications performance is quantified in terms of the output signal to noise ratio (SNR) and the bit error rate (BER) in the soft demodulation output. Fig. 5(a) is the decoding constellation for VS01; Fig. 5(b) is the decoding constellation for VS02; Fig. 5(c) is the decoding constellation for VS03; Fig. 5(d) is the decoding constellation for VS04. Table 1 shows the statistical experimental results. Though there was no sound speed profiler available during this field experiment, one could easily know the sound speed was a
4
X. Han et al. / Applied Acoustics 98 (2015) 1–5 Table 1 The experimental results.
Fig. 4. Channel impulse response detected from different received signals during the SHJ14 experiment at communication distance of 1000 m (a) VS01, the depth of vector sensor is 2.5 m; (b) VS02, the depth of vector sensor is 2.0 m; (a) VS03, the depth of vector sensor is 1.5 m; (a) VS04, the depth of vector sensor is 1.0 m.
slightly positive gradient distribution given that the environment temperature was around zeros degree Celsius. So there would be more acoustic energy in the area near the river surface. It can also be seen from Fig. 5 that the decoding results of VS01 and VS02 are much better than that of VS03 and VS04 which proves the above inference. Another interesting phenomenon is that decoding output SNRs of VS02 and VS03 are much higher than that of VS01 and VS04 respectively. An obvious reason is the difference between
Signals
Output SNR (dB)
BER (%)
VS01 VS02 VS03 VS04
9.90 12.94 9.25 8.75
0.21 (55/26,000) 0 (0/26,000) 0.41 (107/26,000) 0.62 (160/26,000)
their channels. As point C and point B located in the middle of the water column (depicted in Fig. 3), received signals at these two points suffered minor influence from the surface and bottom, so the multipath structures were much simpler. Considering the depth of experimental water column is only 6 m, the decoded results are very reasonable. Fig. 6 shows the decoding constellation using TRM when asynchronously combining the four received signals VS01, VS02, VS03, and VS04 together based on the spatial and temporal diversity. The output SNR = 16.77 dB and BER = 0(0/ 26,000). Several comments are in order. As observed in Figs. 5 and 6, the combination of P, vx, vy has more ability to combat the multipath of UWA channels, so a good performance is achieved even using a single vector sensor. There is a significant improvement in performance when decoding from using a single vector sensor to TRM which asynchronously combines the received signals together. The output SNR has been improved by 6.87 dB, 3.83 dB, 7.52 dB, and 8.02 dB for VS01, VS02, VS03, and VS04 respectively. These results show that multichannel processing could get valuable gains based on spatial and temporal diversity. Khosla and Dowling [18] pointed out that there would be a gain of 10 log M (M = 4 in this
Fig. 5. The decoding constellation of received signal VS01, VS02, VS03, and VS04 using a single vector sensor. (a) VS01: point A, the depth is 2.5 m; (b) VS02: point B, the depth is 2.0 m; (c) VS03: point C, the depth is 1.5 m; (d) VS04: point D, the depth is 1.0 m.
5
X. Han et al. / Applied Acoustics 98 (2015) 1–5
the China Postdoctoral 2013M531015).
Science
Foundation
(Grant
No.
References
Fig. 6. The decoding constellation using TRM when asynchronously combining received signal VS01, VS02, VS03, and VS04 together based on spatial and temporal diversity.
paper), 6.02 dB in theory when decoding using TRM with pressure sensors. The gains here are much higher than that as the combination of p(t) and v(t) could get additional gains. 5. Summary This paper presents a new underwater acoustic communication scheme, which utilizes vector sensor processing and time reversal technology. Single carrier underwater acoustic communication using a vector sensor is investigated based on experimental data collected in shallow water. The same signal was transmitted four times when the vector sensor was deployed to different depths. Decoding results show that even a single channel processing could get robust performance in a highly refractive environment. When asynchronously combining the received signals using TRM based on spatial and temporal diversity, additional 6.87 dB, 3.83 dB, 7.52 dB, and 8.02 dB gains are obtained for VS01, VS02, VS03, and VS04 respectively. Communications using vector sensor with Doppler shifts will be a further research topic in the future. Acknowledgements This research is supported by the National Natural Science Foundation of China (Grant Nos. 51179034, 61471137), Ships Pre-research Support Technology Fund (Grant No. 13J3.1.5), and
[1] Song HC, Howe Bruce M, Brown Michael G, Andrew Rex K. Diversity-based acoustic communication with a glider in deep water. J Acoust Soc Am 2014;135(3):EL1023–6. [2] Ying Xiao, Fu-liang Yin. Blind equalization based on RLS algorithm using adaptive forgetting factor for underwater acoustic channel. China Ocean Eng 2014;28(3):401–8. [3] Huang SH, Tsao J, Yang TC, Cheng SW. Model-based signal subspace channel tracking for correlated underwater acoustic communication channels. IEEE J. Ocean Eng. 2014;39(2):343–56. [4] Leus Geert, van Walree Paul A. Multiband OFDM for covert acoustic communications. IEEE J Sel Areas Commun 2008;26(9):1662–73. [5] Tao Jun, Zheng Yahong Rosa, Xiao Chengshan, Yang Robust TC. MIMO underwater acoustic communications using turbo block decision-feedback equalization. IEEE J Ocean Eng 2010;35(4):948–60. [6] Huang Jianzhong, Zhou Shengli, Huang Jie. Progressive inter-carrier interference equalization for OFDM transmission over time-varying underwater acoustic channels. IEEE J Sel Topic Signal Process 2011;5(8):1524–36. [7] Polprasert Chantri, Ritcey James A, Stojanovic Milica. Capacity of OFDM systems over fading underwater acoustic channels. J Ocean Eng 2011;36(4):514–24. [8] Stojanovic Milica, Catipovic Josko A, Proakis John G. Phase-coherent digital communications for underwater acoustic channels. IEEE J Ocean Eng 1994;19(1):100–11. [9] Xin Zhang, Xiao-Ji Zhang, Xiao-Fei Xing, Li-Wei Jiang. Frequency domain response estimation of underwater acoustic channel and noise estimation for single-carrier frequency domain equalization. Acta Phys Sin 2014;63(19). 194304_1-194304_8. [10] Xia Menglu, Rouseff Daniel, Ritcey James A, Zou Xiang, Chantri Polprasert, Xu Wen. Underwater acoustic communication in a highly refractive environment using SC–FDE 2014;39(3):491–9. [11] Cheng-Bing He, Jian-Guo Huang, Qing-Wei Meng, Qun-Fei Zhang, Wen-Tao Shi. PN-based single carrier block transmission with iterative frequency domain equalization over underwater acoustic channels. Acta Phys Sin 2013;62(23). 234301_1–234301_7. [12] Song HC, Hodgkiss WS. Efficient use of bandwidth for underwater acoustic communication. J Acoust Soc Am 2013;134(2):EL905–8. [13] Shimura T, Ochi H, Song HC. Experimental demonstration of multiuser communication in deep water using time reversal. J Acoust Soc Am 2013;134(4):3223–9. [14] Song HC. Equivalence of adaptive time reversal and least squares for cross talk mitigation. J Acoust Soc Am 2014;135(3):EL154–8. [15] Stojanovic M, Proakis JG, Catipovic JA. Performance of high-rate adaptive equalization on a shallow water acoustic channel. J Acoust Soc Am 1996;100(4):2213–9. [16] Flynn John A, Ritcey James A, Rouseff Daniel, Fox Warren LJ. Multichannel equalization by decision-directed passive phase conjugation: experimental results. IEEE J Ocean Eng 2004;29(3):824–36. [17] Jingwei Yin, Sen Yang, Pengyu Du, Yun Yu, Yang Chen. Code divided multiple access underwater acoustic communication based on active average acoustic intensity. Acta Phys Sin 2012;61(6). 064302_1–064302_7. [18] Khosla Sunny R, Dowling David R. Time-reversing array retrofocusing in noisy environments. J Acoust Soc Am 2001;109(2):538–46.