Autonomous Route Planning of Unmanned Underwater Cluster Based on Reinforcement Learning

The unmanned underwater cluster realizes multi-node cooperative underwater situational awareness using communication networking, navigation control and information perception, which is faced with problems such as weak information connecting, the lack of single node environment and target information. Thereinto, how to solve the autonomous unmanned cluster route planning is very challenging and very important basic problem. Therefore, combining with the Bayesian framework, this paper proposes a method based on reinforcement learning, which aimed at maximization of cluster communication-detecting comprehensive efficiency, and implementing autonomous route planning using the feedback information of unknown obstacles and threats based on multi-node environment - target situational awareness. Through theoretical modeling and simulation analysis, this paper verifies the effectiveness of the proposed method, which can be used to follow-up unmanned underwater cluster technology.