Informed RRT* with improved converging rate by adopting wrapping procedure

Abstract

Wrapping-based informed RRT*, proposed in this paper, combines a size-diminishing procedure, i.e., 'wrapping procedure' with informed RRT*, which samples random path nodes within a hyperellipsoid. The major and minor axes of the hyperellipsoid are determined by the initial and final configurations and current best solution's path cost. Wrapping-based informed RRT* can advance from the first solution acquired by the planner to an improved, feasible solution which can drastically reduce the size of the hyperellipsoid. This leads to much quicker convergence to the optimal value of the path cost, resulting in the minimum action of the robot joints. The algorithm was tested in various environments with different numbers of joint variables and showed much better performance than the existing planners. Furthermore, the wrapping procedure proved to be a comparably insignificant computational burden regardless of the number of dimensions of the configuration space.