Improved RRT algorithm for robotic arm path planning based on reward strategy

Abstract

This paper proposes the Reciprocal Converging Motion (RCM) strategy to address the shortcomings of the basic RRT algorithm in robotic arm path planning, including high randomness, excessive redundant nodes, numerous path corners, and suboptimal path quality. The algorithm enhances the existing RRT method by incorporating a probabilistic sampling strategy, reward mechanism, and collision detection, while leveraging B-spline interpolation curves for goal-oriented path planning. These improvements effectively reduce redundant nodes, enhance search efficiency, and optimize path smoothness. The algorithm is validated through MATLAB-based simulations in both two-dimensional and three-dimensional environments, as well as real-world experimental verification. Experimental results demonstrate that the proposed RCM algorithm reduces redundant node generation by 60% and decreases path search time by an equivalent margin. Furthermore, the application of B-spline smoothing effectively mitigates oscillatory shocks during the robotic arm’s movement along the planned path. These findings highlight the algorithm’s broad applicability and strong robustness in practical implementations.