Lijun Xu

Orthogonal time–frequency space (OTFS) modulation is widely applied to address the nonuniform Doppler shifts issue while still modeling the Doppler effect as different frequency shifts in various paths. This work formulates a practical zero-padding-OTFS system for underwater acoustic (UWA) communications. Taking the impact of time compression/dilation into account, an effective delay–Doppler domain representation is utilized to describe UWA channel response. A basis expansion model (BEM) is employed to capture delay and Doppler multipath features of UWA channel with a weighted combination of basis functions. Leveraging the sparse Bayesian learning (SBL) inference, BEM coefficients can be reconstructed to accomplish the channel estimation. Combined with a decision feedback scheme, the BEM-based SBL turbo receiver is proposed exchanging soft information among channel estimation, equalization, and decoding, achieving coding and iteration gains. The proposed scheme has been verified through the underwater trial data collected from an experiment conducted in the Guishan Island Sea area, Zhuhai, China, in January 2024. Its effectiveness has been evaluated against other benchmarks in terms of bit error rate.

Orthogonal time frequency space (OTFS) modulation has been introduced to underwater acoustic (UWA) communications. However, due to the wideband nature of UWA channels, the Doppler effect cannot be formulated with a uniform frequency offset. In view of this, we pursue a UWA OTFS model and corresponding channel estimation scheme in this paper, considering the Doppler-introduced time compress/dilation effect. By analyzing the characterization of UWA OTFS input-output relationships, the formulation of effective delay-Doppler (DD) domain UWA channel response is provided at first, where we also demonstrate the generation of inter-path interference (IPI) among the DD grids. Further, we show that the effective channel matrix can be formulated by a low-dimensional weighted combination of basis functions based on the basis-expansion-model (BEM) concept, thus constructing the channel estimation as a sparse coefficients recovery problem. To address this issue, we then propose a BEM-based iterative sparse Bayesian learning (SBL) channel estimation algorithm leveraging the framework of SBL and the iterative decision feedback scheme. Finally, simulation results validate the performance advantages of our proposed channel estimation algorithm by comparing it to the typical benchmarks.

Knowledge distillation (KD) is a predominant technique to streamline deep-learning-based recognition models for practical underwater deployments. However, existing KD methods for underwater acoustic target recognition face two problems: 1) the knowledge learning paradigm is not very consistent with the characteristics of underwater acoustics and 2) the complexity of acoustic signals in ocean environments leads to different prediction capacities in teacher and student models. This induces feature misalignment in the knowledge transfer, rendering suboptimal results. To address these problems, we propose a new distillation paradigm, i.e., channel–spatial aligned global knowledge distillation (CSGKD). Considering that the channel features (indicating the loudness of signals) and spatial features (indicating the propagation patterns of signals) in Mel spectrograms are discriminative for acoustic signal recognition, we design the knowledge-transferring scheme from “channel–spatial” aspects for effective feature extraction. Furthermore, CSGKD introduces a global multilayer alignment strategy, where all student layers collectively correspond to a single teacher layer. This allows the student model to dissect acoustic signals at a granular level, thereby capturing intricate patterns and nuances. CSGKD achieves a seamless blend of richness and efficiency, ensuring swift processing while being detail oriented. Extensive experiments on two real-world oceanic data sets confirm the superior performance of CSGKD compared to existing KD methods, i.e., achieving an accuracy (ACC) of 82.37% ($\uparrow$ 2.49% versus 79.88%). Notably, CSGKD showcases an 8.87% improvement in the ACC of the lightweight student model.

This letter presents a joint impulsive noise mitigation and sparse channel estimation algorithm for single-carrier systems in mobile underwater acoustic communication (UAC) to address the challenges posed by impulsive noise interference. The basis expansion model (BEM) is introduced to describe the underwater acoustic time-varying channels and reduce the channel parameters to be estimated. Based on the model, the variational Bayesian inference algorithm is employed to derive an iterative estimation expression for jointly estimating the sparse BEM channel coefficients and the impulsive noise. The proposed algorithm mitigates the impact of impulsive noise, leading to enhanced communication reliability. Simulations and sea trials validate the effectiveness of the proposed algorithm, highlighting its potential for practical applications in mobile underwater acoustic communication systems.

This paper proposes a joint multi-task Bayesian channel estimation and data detection algorithm for Turbo equalization (TEQ) in underwater acoustic (UWA) communication. The joint channel estimation and data detection (JCED) problem is formulated as a multi-task sparse Bayesian learning framework in single carrier (SC) communications. The framework treats the equalized symbols as unknown variables for improving the performance of iterative equalization and leverages temporal correlation in UWA channels by partitioning received symbols into subblocks. Furthermore, a JCED algorithm is derived with variational Bayesian inference. The proposed algorithm was evaluated based on the underwater field data collected during a lake experiment conducted in Qiandao Lake, Zhejiang province, China, in May 2016. The performance of the proposed algorithm has been validated with simulation and experiment results.

This letter addresses the problem of distributed target detection in structured interference and non-homogeneous disturbance. The target signal and interference locate in two linearly independent subspaces with unknown coordinates, while the disturbance is partially homogeneous with unknown covariance matrix. By incorporating the persymmetric structure of received data, we propose a persymmetric subspace detector for distributed target, which includes the detector for point-like target as a special case. Remarkably, the proposed detector is shown to ensure a constant false alarm rate property with respect to both the covariance matrix structure and the power scaling factor. In addition, analytical expressions for the probabilities of false alarm and detection of the special case of the proposed detector for point-like targets are derived, which are verified by Monte Carlo trials. Numerical examples illustrate that the proposed detector can achieve better detection performance as well as anti-interference ability in training-limited situations.

In underwater acoustic time-varying multipath channels, the traditional block Doppler estimation algorithm performs poorly. In this paper an improved Doppler estimation algorithm using virtual time reversal mirror is proposed. The proposed algorithm automatically matches multipath time-varying channels by focusing synchronous head of two adjacent fame-signals in the time domain, then measures the frame length accurately, and then the estimate of the Doppler shift is obtained. Any communication system transmitting signal by frame can employ the algorithm without additional assisting signals, and with acceptable increase in computational cost as well. The results of the sea experiment in Qingdao show the great reliability of this algorithm, and the Doppler shift can be estimated correctly in different types of time-varying multipath channels.