Homotopic Path Planning for an AUV on Maps Improved with Scan Matching

The main contribution of this paper is an application for an Autonomous Underwater Vehicle (AUV) which first improves dead-reckoning navigation through scan matching techniques to build and Occupancy Grid Map (OGM) and then performs path planning to find safe paths towards the goal. The path planning process is guided with homotopy classes, which topologically describe how paths avoid the obstacles. Then, a homotopic path planner called Homotopic Bug (HBug) efficiently computes a feasible path for each homotopy class in the OGM. The application has been tested on an AUV in a controlled unknown scenario.