AUV Path Planning under Ocean Current Based on Reinforcement Learning in Electronic Chart

Abstract

Autonomous underwater vehicles (AUV) are unmanned underwater robots. They are always used to investigate sea environments, oceanography and deep-sea resources autonomously. Navigation of underwater vehicles is a very demanding task, especially in dynamic environment, which has great reflection on ocean current. In order to avoid different risk and to save energy, the path of AUV is usually calculated in the electronic charts before the task is beginning. But in dynamic environment of ocean, the predefined path is not very efficient. So the ocean current should be considered. In this paper, an AUC local path under ocean current is adjusted by Q Learning methods, which is proved in simulations system on the electronic charts.