Hybrid sampling assisted BiRRT for enhanced robotic arm path planning in complex industrial scenarios
Chengwei Hu,Yinhua Liu,Wenzheng Zhao,Yinan Wang
DOI: https://doi.org/10.1007/s10845-024-02541-w
IF: 8.3
2024-12-15
Journal of Intelligent Manufacturing
Abstract:Robotics-aided production system is extensively used in the era of Industrial 4.0, in which the efficiency and quality of robotic arm path planning is a critical factor influencing the quality of the product and the productivity of the entire system. Optimal path planning for robotic arms in complex environments usually encounters challenges such as high computational cost and non-optimal or even infeasible paths. To address these challenges, we propose an improved Bi-directional Rapidly-exploring Random Trees (BiRRT) algorithm based on sampling, extension, and path optimization. More specifically, a hybrid sampling approach consisting of Goal sampling, Quadrant sampling, and Gaussian sampling is incorporated into the BiRRT algorithm to improve sampling performance in complex environments. Moreover, the support vector machine trained with labeled data is used to identify the sampled nodes with collision risks, which speeds up the collision detection compared with the traditional geometric-based method. Lastly, the initial path is further revised and improved by applying quadratic optimization to reduce the length and thus improve production efficiency. Case studies, including simulation experiments and experiments in real practice, have been conducted to evaluate the effectiveness of the proposed method. By comparing with cutting-edge path planning algorithms, our proposed method can efficiently generate high-quality paths for robotic arms and also demonstrate robust performance over different scenarios.
engineering, manufacturing,computer science, artificial intelligence