OPTIMAL CALIBRATION AND IDENTIFICATION OF A 2-DOF PARALLEL MANIPULATOR WITH REDUNDANT ACTUATION

摘要

In order to implement the optimal calibration and identification, excitation trajectories of parallel manipulators have to be designed according to the optimal criterion. The excitation trajectories are formulated by finite Fourier series, and the optimal trajectory parameters are calculated by nonlinear optimization algorithm. Under the optimal excitation trajectory, the differential evolution (DE) algorithm is applied to calibrate the kinematic parameters, and the weighted least square (WLS) method is applied to identify the dynamic parameters. Furthermore, a dynamic control experiment is implemented on an actual 2-DOF parallel manipulator with redundant actuation, and the experimental results indicate that the calibration and identification results based dynamic control can improve the tracking accuracy.