Byung-Kwon Min

Ball screws and linear motion (LM) guides are commonly used in precision feed drives to enhance efficiency and precision. The presliding friction in these components interferes with motion control performance as a disturbance. Friction model-based compensation techniques have been developed to effectively address the nonlinear friction effects. It requires high modeling accuracy for improving compensation control performance. However, ball screws and LM guides generally have different presliding regimes, which makes it difficult to estimate the overall friction of the system. In this article, to solve the issues for presliding friction regime, a presliding regime matching (PRM) method is developed. The PRM minimizes the difference in the presliding regime between components with factors affecting the presliding regime. The effects of the preload and ball rotation speed on the presliding regime change are experimentally discussed in a ball-screw-driven stage. These factors are used to estimate the presliding regime differences. A friction compensation control test is performed on a testbed where the presliding regime difference between the ball screw and LM guide is minimized. The experimental results demonstrate that the proposed method enhances the friction compensation control performance, reducing the peak error by 50.8% with the LuGre model and by 43.8% with the GMS model.

With the increasing use of robots in industry, the use of multi-robot systems, i.e., systems in which multiple robots work together to perform a single task, has become common. In an environment where multiple robots cooperate, the relative position and orientation error between robots must be minimized to improve the accuracy of tasks. In multi-robot systems, accurate kinematic parameters (KP) and base frame transformation (BFT) are required to minimize the position and orientation error between the robots. In this study, a method for KP and BFT identification method that can minimize the position and orientation error of the multi-robot system without using a position measurement device is proposed. A cost function representing the position and orientation error of the multi-robot system was constructed, and the differential evolution algorithm was used to determine the parameters that minimize the cost function. Furthermore, a system consisting of two industrial robots was constructed to verify the effectiveness of the proposed identification method. Experimental results show that the proposed identification method can reduce the relative position error of the dual-robot system by 79% compared to before calibration.