In underwater acoustic (UWA) adaptive communication system, due to time-varying channel, the transmitter often has outdated channel state information (CSI), which results in low efficiency. UWA channels are much more difficult to estimate and predict than terrestrial wireless channels, given the more severe multipath environments with varying propagation speeds in different locations, non-linear propagation paths, several-order higher propagation latency, mobile transceiver and obstacles in the sea, etc. To handle the complexity, this paper proposes an efficient online CSI prediction model for UWA CSI prediction considering the complicated correlationship of UWA channels in both the time and frequency domains. This paper designs a learning model called CsiPreNet, which is an integration of a one-dimensional convolutional neural network (CNN) and a long short term memory (LSTM) network. The performance is compared with the widely used recursive least squares (RLS) predictor, the approximate linear dependency recursive kernel least-squares (ALD-KRLS), and two common conventional deep neural networks (DNN) predictors, i.e., back propagation neural network (BPNN) and LSTM network using the measured data recorded in the South China Sea. To validate the efficacy of prediction, we investigate the prediction of CSI in simulated downlink UWA orthogonal frequency division multiple access (OFDMA) systems. Specifically, the measured UWA channel is used in the OFDMA system. A joint subcarrier-bit-power adaptive allocation scheme is used for resource allocation. To further improve the performance, we develop an offline-online prediction scheme, enabling the prediction results to be more stable. Simulation and experimental results show that the CsiPreNet has superior performance than the existing solutions, thanks to its capability in capturing both the temporal and frequency correlation of the UWA CSIs.

Due to the high cost of sea trials and the variability of sea states, the duration of experiments is usually too short to fully verify underwater acoustic communication (UAC) performance in a real-ocean environment. Moreover, traditional UAC channel simulators also face issues of inaccurate environmental parameters or mismatched statistical models. To tackle these challenges, we propose UACC-GAN, a data-driven stochastic channel simulator for UAC, offering an innovative solution for channel data augmentation. UACC-GAN uses the generative adversarial network model to learn the latent space of the measured channel data set and then maps the random sampling points in this space into a new time-varying impulse response (TVIR). Our simulator is validated using the small-scale WATERMARK data set collected at sea. The results indicate that the generated TVIR has a realistic delay–Doppler spread and can reproduce time-varying delay path characteristics. The cumulative distribution of multiple 0-D properties also proves the realism of the entire distribution of the generated channel data set. In addition, by relying on the continuity of the latent space, UACC-GAN generates channel characteristics with random fluctuations, such as Doppler spectrum shape, delay energy distribution, and tap covariance, which contributes to more diverse communication testing conditions. Finally, we pass frequency-hopping spread spectrum and orthogonal frequency-division multiplexing communication signals through the generated and measured channels. The comparable results of simulated bit error rate (BER) and actual BER underline the value of the UACC-GAN simulator for communication system design and testing.

In this paper, we study the sparse Bayesian learning (SBL) framework for channel estimation in underwater acoustic (UWA) orthogonal frequency-division multiplexing (OFDM) communication systems, which provides a desirable property of preventing structural error with fewer convergence errors for sparse signal reconstruction compared with the compress sensing (CS)-based methods. First, we design a SBL-based channel estimator for block-by-block processing using the channel sparse structure independently in each block. Then, we propose a joint channel model after Doppler compensation for multi-block joint processing, where the delays of the channels for several consecutive blocks are similar and the path gains exhibit temporal correlation, and we denote a temporal correlation coefficient for path gains to evaluate the strength of the correlation. Furthermore, we propose the temporal multiple SBL (TMSBL)-based channel estimator to jointly estimate the channels by taking advantage of the channel coherence between consecutive OFDM blocks. Results of numerical simulation and sea trial demonstrate the effectiveness of the SBL and TMSBL channel estimator algorithms in time-varying UWA channel, which achieve better channel estimation performance and lower bit error rate compared with the existing CS-based methods, such as orthogonal matching pursuit (OMP) and simultaneous OMP, especially the TMSBL estimator achieves the best performance in strong temporal correlated channels and maintains robustness in weak temporal correlated channels.

The Doppler effect severely degrades the orthogonality of orthogonal frequency division multiplexing (OFDM) systems, and is a severe problem in the underwater acoustic (UWA) communications, leading to nonuniform Doppler shifts for all subcarriers. In order to alleviate Doppler effect, resampling technique by oversized fast Fourier transform (FFT) can achieve fast and accurate resampling in frequency domain. Nevertheless, the accuracy of this conventional FFT based approach depends on the FFT size, which in turn increases receiver complexity. In order to overcome this problem, the Doppler compensation model in frequency domain is simplified by Taylor series expansion in this letter. Based on the simplified model, we propose the Doppler compensation algorithm by using FFT and interpolation to achieve higher compensation accuracy. Simulation results demonstrate that the proposed algorithm shows very similar bit error rate (BER) performance with that of Chirp Z-transform (CZT) based method. However, the complexity of our proposed algorithm is about half of CZT based method. Further, our proposed low-complexity Doppler compensation algorithm can also be applied to the sampling frequency offset (SFO) correction in terrestrial OFDM communication systems.

To improve the throughput of underwater acoustic (UWA) networking, the In-band full-duplex (IBFD) communication is one of the most vital pieces of research. The major drawback of IBFD-UWA communication is Self-Interference (SI). This paper presents a digital SI cancellation algorithm for asynchronous IBFD-UWA communication system. We focus on two issues: one is asynchronous communication dissimilar to IBFD radio communication, the other is nonlinear distortion caused by power amplifier (PA). First, we discuss asynchronous IBFD-UWA signal model with the nonlinear distortion of PA. Then, we design a scheme for asynchronous IBFD-UWA communication utilizing the non-overlapping region between SI and intended signal to estimate the nonlinear SI channel. To cancel the nonlinear distortion caused by PA, we propose an Over-Parameterization based Recursive Least Squares (RLS) algorithm (OPRLS) to estimate the nonlinear SI channel. Furthermore, we present the OPRLS with a sparse constraint to estimate the SI channel, which reduces the requirement of the length of the non-overlapping region. Finally, we verify our concept through simulation and the pool experiment. Results demonstrate that the proposed digital SI cancellation scheme can cancel SI efficiently.

The orthogonal matching pursuit (OMP) algorithm has been widely adopted in sparse channel estimation, both in radio and underwater acoustic communications. Traditionally, the accuracy of path delay estimation in an OMP based channel estimator depends on the adopted delay grid. The finer the delay grid, the higher the estimation accuracy, but meanwhile the higher the computational complexity. In this letter, we propose a modified OMP based channel estimation scheme for orthogonal frequency division multiplexing with equispaced pilots, which utilizes the characteristic of the Hermitian inner product between the pilot-compensated frequency observation vector and the columns of the dictionary matrix to provide a closed-form estimate for the path delays. In addition, a two-step scheme facilitated by Goertzel's algorithm is proposed to search the samples in the closed-form estimation to further reduce the computational complexity. Simulation results and complexity analysis demonstrate the advantages of the proposed scheme in both estimation accuracy and computational complexity.