Nonlinear and Adaptive Control of Force and Compliance in Manipulators

Abstract

This paper presents new position-based force and compliance con trol schemes for robot manipulators using nonlinear and adaptive controllers. The proposed controllers are simple and can easily be implemented in the outer feedback loop that encloses the inner Cartesian position control system supplied by the robot manufac turer. In the position-based force control approach, adaptive PID and nonlinear PI controllers are proposed for setpoint tracking in which the controller gains are adjusted on-line as functions of the force error. In the position-based compliance control approach, nonlinear and adaptive PD controllers are developed to achieve a desired behavior at contact, where the controller gains are func tions of the contact force. Dynamic simulation results for a 7-DOF Robotics Research Corporation (RRC) arm are presented to demon strate the efficacy of the adaptive compliance control scheme in executing contact tasks. Experimental results are also presented to illustrate nonlinear force control of the RRC arm. The results demonstrate that the proposed nonlinear PI controller, where the integral gain is a sigmoidal function of the force error, yields supe rior performance compared to afixed-gain PI controller by produc ing a fast force response with a small settling time while preventing oscillations. Finally, the nonlinear PD compliance controller is im plemented on the RRC arm and experimental results are presented. These results demonstrate that the nonlinear PD controller, where the proportional gain is a hyperbolic function of the contact force, exhibits less sensitivity to the surface stiffness than afixed-gain PD controller.