An Improved Bidirectional Path Planning Algorithm with Adaptive Target Gravity for Manipulator Path Planning

To enhance the search efficiency and path quality of sampling-based path planning algorithms in high-dimensional spaces, an improved RRT* algorithm (ATG-RRT*) integrating adaptive target gravity and dynamic step-size strategies was proposed. Based on the traditional RRT*, the algorithm introduces an adaptive target gravity strategy to guide the growth direction of the tree, thereby enhancing its adaptability to complex regions. In addition, a dynamic step-size strategy is incorporated to adjust the expansion distance adaptively according to local environmental information, effectively avoiding path redundancy and the risk of getting trapped in obstacles. To validate the algorithm's performance, simulation experiments were conducted on both 2D and 3D maps with varying obstacle complexities, and comparative analyses were performed against RRT -Connect and standard RRT* algorithms. The experimental results demonstrate that ATG-RRT* outperforms the compared algorithms in terms of path length, planning time, and node utilization, with notably stronger scalability and stability in 3D environments. This method not only improves the overall performance of path planning but also provides a solid theoretical foundation for its application in practical robotic systems.