Path Planning for Unmanned Ground Vehicle In Urban Parking Area

This paper focuses on development of path planning for an Unmanned Ground Vehicle (UGV) during the parking behavior operation in an urban area. The proposed method developed in this paper is the Rapidly-exploring Random Tree (RRT) to search for a collision-free trajectory in an environment with moving and static obstacles. The collision avoidance is based on a geometric search in the transformed vehicle's space. The line segments of the collision free path that is derived from the proposed method is interpolated to yield a smooth and short path. In order to test the strategy a C/C++ simulator incorporating an openGL visualizer was developed. The test vehicle is the Toyota Highlander hybrid vehicle which was developed for the 2007 DARPA Urban Challenge.