In order to meet the needs of large-scale underwater operations, the underwater acoustic communication network emerged, marking a historic moment. At the same time, the development of artificial intelligence has promoted the application of intelligent underwater robots in large-scale underwater operations, and the research on related algorithms has been gradually promoted. Due to the complexity of underwater operations and the difficulty of replacing batteries, the energy efficiency of intelligent underwater robots is particularly important in multi-AUVs data acquisition systems. In view of the energy consumption of multi-AUVs data acquisition systems in water acoustic cluster networks, this paper proposed the AE (A*-Energy) algorithm for multi-AUVs task assignment and path planning. Through the simulation experiment, it was proved that the AE algorithm proposed in this paper can effectively reduce the energy consumption of multi-AUVs data acquisition systems and has good energy efficiency.

Underwater acoustic signals and marine environmental noise are statistically independent of each other, which satisfies the basic assumptions of blind source separation problems. Therefore, blind source separation can be applied to the field of underwater acoustic signal processing. In this paper, an efficient system that combines tensor decomposition with Independent Low-rank Matrix Analysis (ILRMA) is proposed. The optimization system is based on the assumption that spatial matrices can be jointly diagonalized. In this case, the higher-order characteristics of tensors can extract the spatial features of the observed signals and provide excellent initial values for the system. Experimental findings express that the proposed algorithm improves separation accuracy while also achieving quicker convergence.

Automatic modulation recognition plays an important role in many military and civilian applications, including cognitive radio, spectrum sensing, signal surveillance, and interference identification. Due to the powerful ability of deep learning to extract hidden features and perform classification, it can extract highly separative features from massive signal samples. Considering the condition of limited training samples, we propose a semi-supervised learning framework based on Haar timefrequency (HTF) mask data augmentation and the positionalspatial attention (PSA) mechanism. Specifically, the HTF mask is designed to increase data diversity, and the PSA is designed to address the limited receptive field of the convolutional layer and enhance the feature extraction capability of the constructed network. Extensive experimental results obtained on the public RML2016.10a dataset show that the proposed semi-supervised framework utilizes 1 of the given labeled data and reaches a recognition accuracy of 92.09 under 6dB signals.

The time-of-flight (TOF) of ultrasonic flow meters is usually determined by the zero-crossing point extracted from the echo, but the zero-crossing point is not necessarily the starting point of the echo, and the time interval (called interval 2) between them is often ignored. This will result in significant measurement deviation in short distances or when the frequency of the transmitted signal is low. Therefore, this paper proposes a new method for calculating the short-distance ultrasonic TOF based on double-interval. The time interval (called interval 1) for continuous signal transmission is considered at the transmitting end, and the interval 2 is considered at the receiving end. It is also validated that the optimal double intervals remain constant under the specific excitation signal of the transmitting transducer, and the TOF is corrected accordingly. The method proposed in this paper has significantly improved the flow velocity measurement performance compared to the traditional double-threshold (DT) and the method based on intermittent non-coding excitation, with a simulated flow measurement accuracy of 99.974%. Moreover, this paper also improves traditional Kalman filtering for better data preprocessing.