The deep-sea landing vehicle (DSLV) swarm exploration system is a novel platform for the detection of marine mineral resources. A high-precision cooperative localization system with Ultra-Short Baseline (USBL), Doppler Velocity Log (DVL), and electronic compass (EC) plays a vital role in the DSLV swarm exploration system. However, DVL measurements can be seriously interrupted due to the complex operational underwater environment, leading to unstable localization performance. The accuracy of the cooperative localization system could be further degraded by the persistent rubber track slippage during the vehicle’s movement over the soft seabed. In this study, a data-driven cooperative localization algorithm with a velocity prediction model is proposed to improve the positioning accuracy of DSLV under track slippage. First, a velocity prediction model for DVL measurements is constructed using multi-output least squares support vector regression (MLSSVR), and a genetic algorithm (GA) is further employed to optimize the model’s hyperparameters in order to enhance the robustness of the framework. Furthermore, the outputs of MLSSVR are fed into a DSLV position estimation framework based on the Unscented Kalman Filter (UKF) to improve localization accuracy in the presence of DVL failures. To validate the proposed method, the RecurDyn multibody dynamics simulation platform is applied for data synthesis, accounting for both the impact of the soft seabed and real-world motion simulation. The experimental results indicate that during DVL failure, the proposed algorithm can effectively compensate for the cooperative localization errors caused by track slippage, thereby significantly improving the accuracy and reliability of the DSLV cooperative localization system.

Long-time coherent integration is an effective method for dim target detection from heavy sea clutter. To detect dim targets, a novel long-time coherent integration method based on segmented compensation is proposed in this paper. The method models the complex motion of a marine target as the combination of multi-stage uniformly accelerated motions. According to the difference of energy distribution in Doppler frequency domain, this method can suppress sea clutter and detect the regions of interest (ROIs). Using time–frequency domain energy analysis, the potential target can be extracted. After estimating the parameters and segmentation, for the potential targets, the phase compensation factor can be used to eliminate the Doppler frequency modulation caused by the complex motion. Finally, for the compensated signal, long-time coherent integration is performed to realize the target detection and discrimination under low signal-to-clutter ratio. To verify the effectiveness of the proposed method, we apply simulation data and measured CSIR data in the experiments. The results show that the proposed method can integrate the energy of target more effectively than MTD and RFrFT, and the novel method has better detection performance for complex moving targets under low signal-to-clutter ratio situation.

Facing the challenges of scarce annotations in forward-looking sonar image segmentation, this paper proposes a teacher–student network for unsupervised domain adaptation. The proposed model undergoes supervised learning with optical image data to endow the student model with basic segmentation capabilities, using the segment anything model (SAM) as the teacher to generate pseudo-labels in the sonar image domain, thus achieving knowledge transfer without relying on annotated sonar images. An adaptive weighting approach is proposed, which generates a consistency map using predictive consistency in the source domain and the target domain to assess the quality of pseudo-labels. This method dynamically adjusts supervisory strength, preventing incorrect fitting caused by noisy pseudo-labels. In addition, a multi-scale attention module is designed to refine bottleneck features of the U-Net. The effectiveness of the proposed method is validated on a self-built public forward-looking sonar image dataset, achieving a mean intersection over union (IoU) of 40.8% and a mean average precision (mAP) of 70.3%, demonstrating significant improvements over existing typical UDA methods.

Anthropogenic underwater noise is an emerging pollutant, urging a focus on studying the spatial and temporal dynamics of underwater soundscapes for effective environmental protection. This study centers on a bustling harbor enriched by coral ecosystems, employing 5–8 spatially distributed bottom-mounted passive acoustic monitoring (PAM) nodes strategically placed at locations of interest with diverse natural conditions and ship traffic intensities. Focused on the intricate interplay between anthropogenic and natural contributors, including shipping traffic and coral ecosystems, one-third octave band sound pressure levels SPLs were calculated and analyzed at different frequencies, revealing significant variations in levels across different times and locations. Diurnal variations in ambient noise levels highlight the complex relationship between human activities and marine ecosystems. Higher and more variable SPLs, primarily related to shipping traffic, were recorded during the daytime, while lower SPLs, accompanied by increased fish sounds, were observed at night. The collaborative deployment of multiple hydrophones enables an estimation of noise source direction, and multi-hydrophone based principal component analysis (PCA) for dimensionality reduction and subsequent clustering methods were used for noise source classification. This study significantly contributes to the field of underwater noise monitoring, offering a holistic perspective on busy harbor environments.