Improved RRT-Connect Manipulator Path Planning in a Multi-Obstacle Narrow Environment

• Published in: Sensors (Basel, Switzerland) Vol. 25; no. 8; p. 2364
• Main Authors: He, Xueyi, Zhou, Yimin, Liu, Haonan, Shang, Wanfeng
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 08.04.2025; MDPI
• Subjects: adaptive step size; Algorithms; Bias; Costs; Deep learning; Efficiency; Expansion; Heuristic; heuristic sampling strategy; Neural networks; node rejection strategy; Optimization; path planning; Planning; Probability; robotic arm; Robotics; RRT-Connect; Trees; heuristic sampling strategy; node rejection strategy; adaptive step size; path planning; RRT-Connect; robotic arm
• DOI: 10.3390/s25082364
• ISSN: 1424-8220

This paper proposes an improved RRT*-Connect algorithm (IRRT*-Connect) for robotic arm path planning in narrow environments with multiple obstacles. A heuristic sampling strategy is adopted with the integration of the ellipsoidal subset sampling and goal-biased sampling strategies, which can continuously compress the sampling space to enhance the sampling efficiency. During the node expansion process, an adaptive step-size method is introduced to dynamically adjust the step size based on the obstacle information, while a node rejection strategy is used to accelerate the search process so as to generate a near-optimal collision-free path. A pruning optimization strategy is also proposed to eliminate the redundant nodes from the path. Furthermore, a cubic non-uniform B-spline interpolation algorithm is applied to smooth the generated path. Finally, simulation experiments of the IRRT*-Connect algorithm are conducted in Python and ROS, and physical experiments are performed on a UR5 robotic arm. By comparing with the existing algorithms, it is demonstrated that the proposed method can achieve shorter planning times and lower path costs of the manipulator operation.