Impedance estimation for robot contact with uncalibrated environments

Abstract

This paper examines the problem of online estimation of the environment impedance parameters and position during the contact of a robot with an uncalibrated environment. Position controllers are commonly used for free motion, while force controllers are used for constrained motion of robots. However, it is not trivial to set controller parameters for practical contact tasks with unknown environments. In particular, the environmental position is uncalibrated, and it is impossible to determine the contact position through force sensor information because of the presence of a force measurement threshold. To this end, an estimation algorithm based on L-BFGS is proposed to obtain the environmental impedance parameters and position. The estimated environment position is updated with the impedance parameters, and the impedance parameter estimation is directly related to the penetration position. The penetration position is determined by the real position of manipulators and environment surface position. The estimated environmental parameters are used to modify the position and force switching controller parameters; therefore, the accuracy of these parameters can be evaluated by the performance of the modified controller. Simulations and experiments are successfully conducted and validated with the proposed impedance estimation method. The accuracy and effectiveness of the algorithm is verified through the force and trajectory tracking performance in simulations and experiments.