An Iterative Variable Step Error Compensation Approach for Enhancing Robot Trajectory Accuracy

In this paper, an iterative error compensation approach is analyzed and designed to improve trajectory accuracy of industrial robots. Advanced offline programming and robotic machining tasks require high trajectory accuracy of industrial robots, which is significantly affected by parameter inaccuracy, stiffness deformation, backlash, and so on. Such complicated coupling makes it impossible to model each factor separately, thus data-driven direct error compensation is the major solution to improve trajectory accuracy. Previous methods utilize external measurement equipment to obtain tracking error and rely on the error consistency assumption to compensate for the error by modifying the trajectory commands directly. However, such consistency assumption does not hold. Even a minor incremental modification of the initial commands can lead to unexpected influence on the robot's tracking performance, owning to unaccounted factors such as the robot's dynamic stiffness and natural frequency. This paper decomposes trajectory tracking errors based on servo encoder feedback and external laser tracker, and proposes a variable step error compensation scheme to improve tracking performance. Physical experiments are conducted on a standard 6 degree-of-freedom industrial robot to validate the effectiveness of the proposed method. The results of various error compensation approaches are analyzed to draw a possible error correlation model, in order to account for the error inconsistency in the compensation process.