Obstacle Avoidance Trajectory Planning for Truss-Type Coal Gangue Sorting Robot

Coal gangue sorting is essential for improving coal quality and utilizing solid waste. In practical applications, robotic arms must not only complete target grasping but also maintain smooth trajectories and effective obstacle avoidance. Traditional RRT* algorithms face challenges such as long computation times and overly extended paths. This paper proposes an improved RRT* algorithm combining the bounding box method and Artificial Potential Field to enhance obstacle avoidance. A probabilistic bias strategy is in-troduced to maintain randomness while guiding path growth, and cubic B-spline interpolation smooths the trajectory. Experiments comparing the improved RRT* algorithm with the traditional RRT algorithm show that the proposed method is more suitable for coal gangue sorting. Adding target region bias reduced the average runtime by 10.22%, and under this strategy, the improved algorithm achieved shorter path lengths and a 35.89% reduction in runtime. The results demonstrate that the proposed algorithm effectively addresses obstacle avoidance and performs better in coal gangue sorting environments, with shorter paths and faster runtimes.