Choosing Measurement Configurations for Kinematic Calibration of Cable-Driven Parallel Robots

Abstract

For kinematic calibration of cable-driven parallel robots (CDPRs), many kinematic calibration methods have been proposed in the last decade. However, the established methods are mainly based on calibration with different models, whose robustness of observability index with respect to (w.r.t) sensor noise is not investigated. The observability index of the robot configurations can be maximized and the calibration accuracy will be improved by adequately choosing the measurement configurations. This paper proposes to calibrate the kinematic calibration of CDPRs by choosing measurement configurations, which can optimize the observability index and increase the calibration accuracy. The kinematic calibration model is formulated, and then a calibration method is designed to calculate the transformation matrix between the measuring-coordinate frame and the global coordinate frame. With these models, an efficient configuration choosing algorithm is proposed by using the DETMAX algorithm and the Tabu search algorithm. The proposed algorithm was validated by simulation experiments under several observability indices, and the results indicate that the robustness of the calibration method can be improved significantly by choosing configurations. With the theoretical geometric parameters and the selected measurement configurations, calibration experiments were performed on a redundant CDPR.