Research on Lateral Line Array Density Optimization and Target Motion Perception Algorithm

The lateral line sensing principle provides a new idea for underwater target perception and navigation of vehicles in complex flow field environments. However, it is difficult to construct a complex flow field analytical model, which leads to many obstacles in the research of accurately perceiving underwater moving targets. This paper uses a high-fidelity simulation model of the pressure field change of the lateral line array caused by the moving source to evaluate the perception sensitivity and discreteness of the lateral line array based on the moving source's moving speed and perception distance. The simulation results show that the sensor density of the lateral line array is closely related to the performance of the array data acquisition. The optimal resolution of the lateral line array is 50 pcs/m. At this resolution, the pressure distribution generated by the moving source has high recognition and can be used for the training and prediction of deep learning models. At the same time, a large-scale time series data set is constructed based on the high-fidelity simulation experiment of the moving source's multi-angle and multi-speed conditions, and the position of the moving source is further predicted based on the RNN-LSTM and ConvLSTM network model. The results show that the average recognition error of the underwater linear moving target position is within 7cm. Based on the array design of the lateral line sensor and the prediction of the deep learning model, this paper provides a new method for underwater target perception and positioning of vehicles in complex flow fields.