A Novel Positioning Accuracy Improvement Method for Polishing Robot Based on Levenberg–Marquardt and Opposition-based Learning Squirrel Search Algorithm

Achieving high-precision manufacturing of optical components requires improving the absolute positioning accuracy of the robot to the highest possible level. Identifying the robot's kinematic parameters and compensating for kinematic errors are effective methods for improving the robot's positioning accuracy. This paper proposes a hybrid algorithm that combines the Levenberg–Marquardt algorithm and an opposition-based learning squirrel search algorithm to identify the kinematic parameters of a polishing robot. Firstly, the Levenberg–Marquardt algorithm is utilized to solve the suboptimal values of kinematic parameter deviations. Secondly, an opposition-based learning strategy is integrated into the standard squirrel search algorithm to increase the diversity of the population and prevent the population from getting stuck in local optima. The suboptimal values obtained by the Levenberg–Marquardt algorithm are subsequently used as the central values to generate the initial population for the opposition-based learning squirrel search algorithm, which helps identify more accurate kinematic parameter deviations. Ultimately, the kinematic parameters of the robot are effectively calibration. The calibration experimental results showed that the proposed method achieved a high level of calibration accuracy, resulting in a 62.61% improvement in absolute positioning error compared to before calibration. Offline machining experiments have validated the effectiveness of LM-OBLSSA in reducing deviations in the dwell points of optical components during the machining process.