Research on underwater target tracking method based on IPSO-LSTM

When tracking underwater targets, due to the limited performance of detection equipment and complex environment interference, the target data is lost in the transmission process, which affects the accuracy of target trajectory prediction. The traditional target trajectory prediction method for establishing mathematical model has some limitations and difficulty, which can not meet the requirements of real-time and accurate target trajectory prediction in underwater field. An IPSOLSTM trajectory prediction model for underwater targets is proposed. In this method, the motion characteristics of underwater targets are analyzed, and the learning rate of LSTM is optimized by optimizing the inertia weight of particle swarm and improving the learning factor. The real trajectory data of Autonomous Underwater Vehicle were used for experimental and comparative analysis. The results show that IPSO-LSTM combined with PSO has higher prediction accuracy and better fitting effect than LSTM and PSO-LSTM. The proposed model provides a new idea for underwater moving target tracking, which has better accuracy and stability.