Modified Rapidly-exploring Random Tree with A-star Heuristic for Static Obstacle Avoidance in Robotic Arm

Efficient motion planning is vital for robotic arms operating in environments with static obstacles. The traditional rapidly-exploring random tree path-planning algorithm has been extensively used for robotic arm motion planning. Despite its capability to explore high-dimensional spaces, the rapidly-exploring random tree often generates suboptimal paths due to its nature in random exploration of nodes. This paper proposes a modified rapidly-exploring random tree with A-star heuristic that integrates the goal-oriented searching of A-star and Manhattan distance heuristic into the original algorithm to prioritize axis-aligned movements. A comparative analysis was conducted between rapidly-exploring random tree, rapidly-exploring random tree with A-star heuristic, and the modified rapidly-exploring random tree with A-star heuristic through simulation. The simulation environment included a Universal robot model, two vision sensors, and a fixed cube obstacle. Two predefined motions, namely; motion on x-y plane, and motion on 3D space were programmed to evaluate the performance of the considered algorithms. The performance metrics included total distance travelled and motion completion time. Results show that the proposed algorithm significantly reduced the motion completion time by at least 35% while maintaining a smooth path, making it suitable for the real-time robotic arm applications.