Virtual musculoskeletal control model with a spindle-like fuzzy algorithm for robotic compliance

Robot manipulators should comply with the environment when working with humans. However, compliance control is a difficult problem for robots and thus it was investigated in the present study. We propose a virtual musculoskeletal control model from the perspective of bionics to facilitate the compliance of the robotic manipulator. The musculoskeletal system of the human forearm is simplified as a closed-loop control system with three parts: the central nervous system, muscles, and spindle. A mathematical model is deduced and integrated with the dynamic model of the robot manipulator. The control model also comprises three parts: the first part compensates for the Coriolis force and gravity; the second part provides stiffness to regulate the deviation; and the third part imitates the feedback from the spindle to comply with the environment. A fuzzy controller is designed based on the muscle and spindle model to obtain spindle-like feedback. Our simulation results demonstrate that the spindle-like fuzzy algorithm can adapt to the environmental constraints by imitating the function of the neural muscular system. This virtual musculoskeletal methodology allows accurate path control, but it also ensures the compliance of the robotic manipulator. These characteristics are helpful for allowing robot manipulators to cooperate with humans.