Path Planning of Soccer Robot Based on Improved Heuristic RRT* Algorithm

Abstract

For problems solving the soccer robot path planning in the RoboCup standard platform league, an improved algorithm based on Informed Rapidly-exploring random tree* (RRT*) is proposed. A series of improvements have been made to solve the poor purpose and low efficiency in searching for the initial solution in the Informed RRT* algorithm. Firstly, by estimating the initial states and adopting proper inflation strategy, the heuristic sampling subset is constructed in advance to increase the time proportion of constrained sampling. Secondly, the synchronous growth of bidirectional tree is used to replace the traditional alternating growth to simplify the planning task of single random tree, and a special stage is introduced to accelerate the connection. Finally, we adopt the hybrid cache mechanism to reuse the existing nodes, so as to reduce the calculation scale. A series of experimental comparisons show that the improved algorithm has better real-time performance and efficiency.