Due to the stable propagation of magnetic anomaly signal in ocean and air, magnetic anomaly detection technology has become an effective means for non-acoustic detection. Deep learning, particularly convolutional neural networks (CNN), has made significant progress in underwater signal detection. CNN has become an important tool in this field due to its advantages in handling large-scale data, automatic feature extraction, and pattern recognition. However, magnetic anomaly signal is often accompanied by noise, and its spatial and temporal characteristics can vary significantly, presenting challenges for model training. Moreover, effective preprocessing of the dataset is crucial. To address these issues, this paper proposes a neural network method based on orthogonal basis function (OBF) feature detection. This method combines OBF preprocessing with a combined training approach using CNN and long short-term memory networks, which effectively reduces the false alarm rate commonly associated with traditional methods. Consequently, the detection rate is substantially improved, leading to enhanced overall detection performance for magnetic anomaly signals.

Underwater sonar target recognition is crucial in fields such as national defense, navigation, and environmental monitoring. However, it faces issues such as the complex characteristics of ship-radiated noise, imbalanced data distribution, non-stationarity, and bottlenecks of existing technologies. This paper proposes the MultiFuseNet-AID network, aiming to address these challenges. The network includes the TriFusion block module, the novel lightweight attention residual network (NLARN), the long- and short-term attention (LSTA) module, and the Mamba module. Through the TriFusion block module, the original, differential, and cumulative signals are processed in parallel, and features such as MFCC, CQT, and Fbank are fused to achieve deep multi-domain feature fusion, thereby enhancing the signal representation ability. The NLARN was optimized based on the ResNet architecture, with the SE attention mechanism embedded. Combined with the long- and short-term attention (LSTA) and the Mamba module, it could capture long-sequence dependencies with an O(N) complexity, completing the optimization of lightweight long sequence modeling. At the same time, with the help of feature fusion, and layer normalization and residual connections of the Mamba module, the adaptability of the model in complex scenarios with imbalanced data and strong noise was enhanced. On the DeepShip and ShipsEar datasets, the recognition rates of this model reached 98.39% and 99.77%, respectively. The number of parameters and the number of floating point operations were significantly lower than those of classical models, and it showed good stability and generalization ability under different sample label ratios. The research shows that the MultiFuseNet-AID network effectively broke through the bottlenecks of existing technologies. However, there is still room for improvement in terms of adaptability to extreme underwater environments, training efficiency, and adaptability to ultra-small devices. It provides a new direction for the development of underwater sonar target recognition technology.

This paper proposes a complex fast recursive least-squares (FRLS) channel-estimation algorithm for single-carrier electromagnetic (EM) communications across the seawater–air interface, where severe attenuation and multipath cause strong SNR fluctuations. By redesigning the input-data structure and using forward–backward joint estimation, FRLS reduces the per-iteration complexity from the quadratic cost of classical RLS to a linear form (14L + 20 operations per iteration, where L is the channel length). Simulations under representative one- to three-path channels show that FRLS achieves the lowest steady-state mean-square deviation (MSD) at low SNR, outperforming LMS, IPNLMS, RLS, and PRLS. Offshore experiments further validate the practicality: after MMSE equalization, FRLS yields higher OSNR and improves the BER distribution, demonstrating an effective accuracy–complexity trade-off for hardware-constrained cross-medium EM links.