Adaptive Real-Time Predictive Compensation Control For 6-Dof Serial Arc Welding Manipulator

Wang Xuanyin,Ding Yuanming

DOI: https://doi.org/10.3901/CJME.2010.03.361

IF: 2.964

2010-01-01

Chinese Journal of Mechanical Engineering

Abstract:Because of long driving chain and great system load inertia, the serial manipulator has a serious time delay problem which leads to significant real-time tracking control errors and damages the welding quality finally. In order to solve the time delay problem and enhance the welding quality, an adaptive real-time predictive compensation control(ARTPCC) is presented in this paper. The ARTPCC technique combines offline identification and online compensation. Based on the neural network system identification technique, the ARTPCC technique identifies the dynamic joint model of the 6-DOF serial arc welding manipulator offline. With the identified dynamic joint model, the ARTPCC technique predicts and compensates the tracking error online using the adaptive friction compensation technique. The ARTPCC technique is proposed in detail in this paper and applied in the real-time tracking control experiment of the 6-DOF serial arc welding manipulator. The tracking control experiment results of the end-effector reference point of the manipulator show that the presented control technique reduces the tracking error, enhances the system response and tracking accuracy efficiently. Meanwhile, the welding experiment results show that the welding seam turns more continuous, uniform and smooth after using the ARTPCC technique. With the ARTPCC technique, the welding quality of the 6-DOF serial arc welding manipulator is highly improved.