Real Time Path Planning of Underwater Robots in Unknown Environment

Abstract

This paper presents a real time path planning scheme for underwater vehicles operating in completely unknown environments. Due to deficiency of necessary information, most of graph generation and corresponding graph searching methods proposed so far cannot be directly applicable. In this paper, a kind of visibility graph called RBVG (rubber band visibility graph) is introduced to generate roadmap. This graph is generated under the assumption that there isn't any obstacle outside of the vehicle's sonar coverage area (here the coverage area includes all of the past covered area). Since this graph might be different from the real one, we call it virtual RBVG. With this graph, we apply an A* algorithm to plan a kind of optimal path. Unlike to the mobile robot case where Cartesian distance is usually taken as heuristic cost function, in the water, various aspects of criteria (e.g. energy consumption and mission duration) should be carefully considered. Some simulation studies are also carried out to demonstrate the effectiveness of proposed path planning scheme.