Joint localisation and tracking for autonomous underwater vehicle: a reinforcement learning‐based approach

Abstract

This study is concerned with a joint localisation and tracking problem for autonomous underwater vehicle (AUV), subject to asynchronous clock and stratification effect in cyber channels as well as the model disturbances in physical channels. The authors first construct an integrated state and clock model, which allows the co-design of communication and control strategies. Then, an asynchronous localisation algorithm is developed to estimate the position of AUV, where the asynchronous clock and the stratification effect are both considered. With the estimated position information, a reinforcement learning based tracking controller is developed for the AUV to track the reference point. Particularly, the multivariate probabilistic collocation method is adopted to evaluate the model uncertainty. Moreover, the convergence analyses for the localisation algorithm and tracking controller are also given. Finally, simulation results are presented to show the effectiveness of the proposed method. It is demonstrated that the communication energy consumption and tracking error can be significantly reduced, through the co-design of localisation and tracking strategies.