2022/01/01 by Sotiris Makris, Emmanouil Kampourakis, Dionisis Andronas · 1 citation
Engineering · Computer Science · #Robot Manipulation and Learning #Robotic Path Planning Algorithms #Teleoperation and Haptic Systems
paper · doi:10.1016/j.cirp.2022.04.048
Despite extensive automation in multiple industrial sectors, manufacturing operations involving deformable objects are mostly performed manually. Challenges originating from flexible objects’ dynamic distortion underline handicaps in robot cognition and dexterity. This paper presents a model-based motion planner for deformable object co-manipulation. The developed closed-loop controlling framework interprets manipulation inputs into appropriate handling actions by simulating fabric's distortion through a mass-spring model. The planner incorporates tools for rapid system commissioning and reconfiguration, grasping point planning, and monitoring of human actions. Inspired by automotive composite industry, two experimental setups are used for validating the system's performance during translational and rotational co-manipulation.