Path Planning for Robot with a Parallel Sampling RRT Algorithm and Trajectory Optimization

The rapidly-exploring random tree (RRT) algorithm is widely used in robot path planning due to inaccurate modeling obstacles in the workspace. However, the RRT algorithm also suffers from excessive invalid sampling points and low planning efficiency resulting from random sampling. To address these issues, a parallel heuristic sampling and bidirectional guidance RRT algorithm (MHBi-RRT) is proposed. First, a parallel heuristic sampling strategy is designed, where multiple points are randomly sampled, and an evaluation function incorporating distance and angle factors is constructed to select the optimal sampling node for expansion. Then, a bidirectional guidance mechanism is established to leverage information from newly generated nodes, accelerating the merging process of the two trees. The expansion step size is determined based on the surrounding environment of the nodes, ensuring the generation of an initial collision-free path. Subsequently, we propose a rewiring trajectory optimization method, in which the initial path is optimized in terms of length and safety. An interpolation B-spline path smoother is presented for easier tracking control of the robot. Finally, the effectiveness of the proposed method in addressing the issues within the RRT is verified by experiments conducted in various map. Furthermore, compared with baseline algorithms, our method reveal the superior planning efficiency and path quality, significantly improving the stability of robot path planning.