Benchmarking of Robot Arm Motion Planning in Cluttered Environments

Abstract

Motion planning is essential for robotic automation across various industries. However, generalizing research outcomes has been challenging due to the narrow focus of previous work on a specific robot arm system. Here, we take a broader approach by exploring the combinations of three popular robot arm systems, three levels of clutterness in the environment, and twelve popular motion planners. To conduct the necessary performance analysis, we employ Motionbenchmaker tool and introduce a sensitivity metric. Our approach is structured and accessible, enabling the identification of the best-performing planner-robotic arm combinations. We find that the LBKPIECE, RRTConnect, and BKPIECE planners with Franka and UR5 offers the best balance of effectiveness and robustness. More generally, our results help researchers and practitioners make informed decisions when selecting robotic arms and motion planners, for use in environments with different degrees of clutterness.