Simultaneous identification of geometric parameters and structural stiffness of serial robots composing flexible links

Abstract

It has been a formidable challenge to accurately predict the pose of serial robots composing flexible links due to the difficulties in simultaneously considering geometric errors and link flexibility. This paper presents an approach for simultaneously identifying kinematics and stiffness parameters in the flexible serial robot. A compliance equivalent method is employed to model the flexibility in the joints and links nonlinearly. Based on this, the kinetostatics model of those serial robots can be established analytically. By approximating the force-deflection characteristics of the flexible links, a unified error model containing both geometric and compliance parameters is derived. Moreover, an adaptive total-least-square method is proposed for the identification considering constraints on the compliance coefficient. To verify the effectiveness of the proposed method, simulations on a single flexible link, a serial robot, and experiments on a 3-DOF planar serial manipulator are carried out. The results show that the residual errors of the manipulator can be significantly reduced by identifying the kinematics and stiffness parameters using the proposed approach.