Research on Robot Path Planning Based on Multi-Strategy Improved RRT* Algorithm

Abstract

To tackle challenges like route redundancy, prolonged computation duration, and convoluted paths within the RRT* algorithm for robot path planning, this research proposes a multi-strategy enhanced APF-RRT* algorithm termed MSAP-RRT*. In order to decrease the number of sampling points and increase the effectiveness of path planning, this method first combines the RRT* algorithm with a dynamic target-biased sampling strategy. Second, to improve the repulsive potential field function in the artificial potential field, obstacle density and the target distance adjustment factor are contained. This enables real-time adjustment of the repulsive force magnitude to improve the quality of path planning. To speed up convergence and cut down on running time, an adaptive step size method based on obstacle density and target distance adjustment factor is then presented. Lastly, cubic B-spline curves and greedy pruning are combined to optimize the original path, reducing its length and enhancing its quality. Simulation comparative analyses conducted in typical scenarios show that the MSAP-RRT* algorithm reduces the average path length by 6.49%, the average runtime by 82.10%, the average number of nodes by 90.70%, and the average number of iterations by 85.20% when compared to the improved APF-RRT* algorithm. Robots are strongly guaranteed to create safe and effective routes in areas with lots of obstacles according to the MSAP-RRT* algorithm.