Chemical Plume Tracing using an AUV based on POMDP Source Mapping and A-star Path Planning

Abstract

This paper presents an engineering-based chemical plume tracing (CPT) method for using on an autonomous under-water vehicle (AUV) to locate a chemical source in an underwater environment with obstacles. Fundamental steps of the proposed method are twofold. Firstly, the estimated source location is obtained by a source likelihood map, which is generated based on a partially observable Markov decision process (POMDP). Secondly, after the estimated source location is determined, the A-star path planning algorithm is used to generate the shortest path toward the target while avoiding obstacles. Simulation results validate the proposed method in environments with either laminar or turbulent flow conditions. Comparing with other chemical source mapping algorithms, such as the hidden Markov model (HMM) based method, the POMDP-based source mapping algorithm converges to the correct chemical source location in a faster rate. Besides, the A-star path planning algorithm enables the AUV to avoid obstacles and the local minima issue in the traditional path planning algorithm, such as the artificial potential field.