Calibration of a (6+3)-Dof Kinematically Redundant Parallel Mechanism Based on Hybrid Identification Algorithm and Two-Step Compensation Strategy

Abstract

To improve both the absolute accuracy and pose repeatability of a (6+3)-DOF kinematically redundant parallel mechanism (KRPM), a novel kinematic calibration method based on hybrid identification algorithm and two-step compensation strategy is proposed. Firstly, the error model is established, and the joint clearances are treated equivalently. On this basis, the global sensitivity index (GSI) is introduced to analyze the influence of parameter errors. Subsequently, the absolute accuracy and pose repeatability are evaluated with the quasi-Monte Carlo simulation method (QMCSM) and the three-dimensional ellipsoidal error probability method (ESEPM). Then, a hybrid identification algorithm combining Levenberg-Marquardt (LM) algorithm and adaptive differential evolution (ADE) algorithm is proposed to ensure the identification speed and accuracy. Finally, the error compensation is carried out by two steps. The experimental results show that based on the novel method, the absolute accuracy and pose repeatability of the KRPM are improved by at least 52.8% and 12.5% respectively.

Keywords: Redundant parallel mechanism, Kinematic calibration, Accuracy evaluation, Identification algorithm, Error compensation