Deep Learning Prediction of Time-Varying Underwater Acoustic Channel Based on LSTM with Attention Mechanism

Abstract

This paper investigates the channel prediction algorithm of the time-varying channels in underwater acoustic (UWA) communication systems using the long short-term memory (LSTM) model with the attention mechanism. AttLstmPreNet is a deep learning model that combines an attention mechanism with LSTM-type models to capture temporal information with different scales from historical UWA channels. The attention mechanism is used to capture sparsity in the time-delay scales and coherence in the gep-time scale under the LSTM framework. The soft attention mechanism is introduced before the LSTM to support the model to focus on the features of input sequences and help improve the learning capacity of the proposed model. The performance of the proposed model is validated using different simulation time-varying UWA channels. Compared with the adaptive channel predictors and the plain LSTM model, the proposed model is better in terms of channel prediction accuracy.

Keywords

Long short-term memory (LSTM) / Attention mechanism / Underwater acoustic communication / Underwater acoustic channel / Channel prediction

References

[1] Bahdanau D, Cho K, Bengio Y (2016) Neural machine translation by jointly learning to align and translate. arXiv:1409.0473 [cs, stat]. Available from http://arxiv.org/abs/1409.0473 [Accessed March 29, 2022]

[2] Fuxjaeger AW, Iltis RA. Adaptive parameter estimation using parallel Kalman filtering for spread spectrum code and doppler tracking. IEEE Transactions on Communications, 1994, 42(6): 2227-2230

[3] Greff K, Srivastava RK, Koutnik J, Steunebrink BR, Schmidhuber J. LSTM: a search space odyssey. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(10): 2222-2232

[4] Huang SH, Yang TC, Huang CF. Multipath correlations in underwater acoustic communication channels. The Journal of the Acoustical Society of America, 2013, 133(4): 21802190

[5] Jiang W, Diamant R (2023) Long-range underwater acoustic channel estimation. IEEE Transactions on Wireless Communications, Early Access. DOI: https://doi.org/10.1109/TWC.2023.3241230

[6] Jiang W, Tong F, Zhu Z. Exploiting rapidly time-varying sparsity for underwater acoustic communication. IEEE Transactions on Vehicular Technology, 2022, 71(9): 9721-9734

[7] Lin N, Sun H, Cheng E, Qi J, Kuai X, Yan J. Prediction based sparse channel estimation for underwater acoustic OFDM. Applied Acoustics, 2015, 96: 94-100

[8] Liu L, Cai L, Ma L, Qiao G. Channel state information prediction for adaptive underwater acoustic downlink OFDMA system: deep neural networks based approach. IEEE Transactions on Vehicular Technology, 2021, 70(9): 9063-9076

[9] Ma L, Xiao F, Li M (2019) Research on time-varying sparse channel prediction algorithm in underwater acoustic channels. 3rd International Conference on Electronic Information Technology and Computer Engineering (EITCE), 2014–2018

[10] Mnih V, Heess N, Graves A, Kavukcuoglu K (2014) Recurrent models of visual attention. Proceedings of the 27th International Conference on Neural Information Processing Systems, 2204–2212

[11] Nadakuditi R, Preisig JC. A channel subspace post-filtering approach to adaptive least-squares estimation. IEEE Transactions on Signal Processing, 2004, 52(7): 1901-1914

[12] Niu Z, Zhong G, Yu H. A review on the attention mechanism of deep learning. Neurocomputing, 2021, 452: 48-62

[13] Oliveira TCA, Lin YT, Porter MB. Underwater sound propagation modeling in a complex shallow water environment. Frontiers in Marine Science, 2021, 8: 751327

[14] Qarabaqi P, Stojanovic M. Statistical characterization and computationally efficient modeling of a class of underwater acoustic communication channels. IEEE Journal of Oceanic Engineering, 2013, 38(4): 701-717

[15] Radosevic A, Duman TM, Proakis JG, Stojanovic M (2011) Channel prediction for adaptive modulation in underwater acoustic communications. OCEANS 2011 IEEE-Spain, 1–5. DOI: https://doi.org/10.1109/Oceans-Spain.2011.6003438

[16] Song A, Stojanovic M, Chitre M. Editorial underwater acoustic communications: where we stand and what is next? IEEE Journal of Oceanic Engineering, 2019, 44(1): 1-6

[17] Tu XB, Xu XM, Song AJ. Frequency-domain decision feedback equalization for single-carrier transmissions in fast time-varying underwater acoustic channels. IEEE Journal of Oceanic Engineering, 2021, 46(2): 704-716

[18] Xu K, Lei J, Kiros R, Cho K, Courville A, Salakhutdinov R, Zemel RS, Bengio Y (2015) Show, attend and tell: neural image caption generation with visual attention. Proceedings of the 32nd International Conference on Machine Learning, 2048–2057

[19] Yang M, Li XK, Yang Y, Meng XX. Characteristic analysis of underwater acoustic scattering echoes in the wavelet transform domain. Journal of Marine Science and Application, 2017, 16(1): 93-101

[20] Yang TC. Properties of underwater acoustic communication channels in shallow water. The Journal of the Acoustical Society of America, 2012, 131(1): 129-145

[21] Zhang Y, Venkatesan R, Dobre OA, Li C. Efficient estimation and prediction for sparse time-varying underwater acoustic channels. IEEE Journal of Oceanic Engineering, 2020, 45(3): 1112-1125

[22] Zhang Z, Hou M, Zhang F, Edwards CR (2019) An LSTM based Kalman filter for spatio-temporal ocean currents assimilation. WUWNET’19: International Conference on Underwater Networks & Systems, Atlanta, 1–7

[23] Zhang ZQ, Al-Abri S, Wu WC, Zhang FM (2020b) Level curve tracking without localization enabled by recurrent neural networks. 5th International Conference on Automation, Control and Robotics Engineering (CACRE), 759–763

[24] Zhou Y, Song A, Tong F. Underwater acoustic channel characteristics and communication performance at 85 kHz. The Journal of the Acoustical Society of America, 2017, 142(4): EL350-EL355

[25] Zhou YH, Cao XL, Tong F. Acoustic MIMO communications in a very shallow water channel. Journal of Marine Science and Application, 2015, 14(4): 434-439

[26] Zhu Z, Tong F, Jiang W, Zhang F, Zhang Z (2021) Evaluating underwater acoustics sensor network based on sparse LMS algorithm driven physical layer. WUWNet’21: The 15th International Conference on Underwater Networks & Systems, Shenzhen, 1–8.

[27] Zhu Z, Tong F, Zhou Y, Wu F. Dual parameters optimization lp-LMS for estimating underwater acoustic channel with uncertain sparsity. Applied Acoustics, 2023, 202: 109150