Maxwell Model-Based Null Space Compliance Control in the Task-Priority Framework for Redundant Manipulators

• Published in: IEEE access Vol. 8; pp. 35892 - 35904
• Main Authors: Fu, Le, Zhao, Jie
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Collaboration; Compliance; Control methods; Deformable models; Force; Manipulators; Maxwell model; Multi-prioritized framework; null space compliance control; plastic deformation; Redundancy; Robot arms; Robots; Safety; Task analysis
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2020.2975125

Improving robot performance while simultaneously ensuring compliance and human safety has been appealing for a long time and remains a challenge. To this end, we propose a new approach for kinematic control of redundant manipulators to deal with multiple prioritized tasks and at the same time produce a novel compliance behavior in the null space of main task. Different from typical Voigt model based compliance control methods, the compliance control approach proposed in this paper is based on Maxwell model, which shows superior performance on impact absorption and contact reducing than its counterparts. In comparison with typical compliance control, the annoying and even harmful return force is removed and human comfortableness can be improved consequently. Besides, this novel compliance control method is implemented in the null space of higher priority task without disturbing the main end-effector task. The effectiveness of our approach has been practically evaluated and verified in experiments on a 7 DOF redundant collaborative robot manipulator. A novel cushion-like and plastic deformed whole body compliance has been realized while the continuity and quality of robot main task have been maintained. Promisingly, the approaches proposed in this paper are general and manipulator independent, which can also be applied to other emerging collaborative and even hyper redundant manipulators. Besides, research results of this work can also potentially inspire other robotic research related with human robot friendly interaction and collaboration.