A Digital Twin-Based Path Planning Method for Bending Robot Feeding and Unloading

During bending operations, the narrow space between the upper and lower dies of a bending machine and the ever-changing shapes of the workpieces make them highly susceptible to collisions with the bending tool. Automatic path planning algorithms are crucial when using robot-assisted bending. This paper proposes an improved Synchronized Biased Greedy RRT -Connect algorithm (Improved SBG-RRT -Connect) for planning a bending robot's feeding and unloading paths. The algorithm addresses the issues of the standard RRT -Connect being easily attracted to low-value regions and having unstable convergence speed by introducing a synchronized biased greedy strategy, adaptive ellipsoidal region sampling, and an adaptive node growth strategy. Compared to the unimproved algorithm, the number of iterations and search time are optimized by 51.91 % and 67.6%, respectively, while achieving less costly paths. Additionally, the algorithm can be pre-tested on the digital twin platform designed in this study, allowing for further control of the physical robot through the virtual system. Experimental results demonstrate that the robot completes collision-free feeding and unloading operations, verifying the algorithm's feasibility.