Decentralized Feedback Design for a Compliant Robot Arm

Enhancing safety during interaction with environment and improved force control has led to design of highly compliant robotic arms. However, due to introduction of passive compliance in series with actuators the links' interactions and coupling effects become much more important. In this paper a direct decentralized approach for designing PD-PID gains, given the dynamic multivariable model of the compliant robot arm, is proposed. The proposed method is based on Linear Matrix Inequality (LMI) formulation with full state feedback in discrete time that automatically designs the decentralized gains for all the four joints of the robot in one shot. Experimental results for a four Degree of Freedom (DoF) compliant robot arm are provided to illustrate the effectiveness of this method.