Multi-Objective Admittance Control: An LMI-Based Method

Abstract

Robust and stable admittance control is critically important in physical human-robot interaction. However, there are inherent structural flaws for the conventional admittance control, where tradeoff should be made among admittance performance (accuracy and passivity) and robustness. This paper addresses the dilemma for the multi-objective optimization problem in admittance control. Firstly, a complementary admittance control framework is proposed with decoupled design freedoms of admittance performance and robustness. Then, LMI-based optimization algorithms are developed to find the controller gains satisfying respective constraints. Finally, simulations are conducted to show the efficacy of the proposed method.