Abstract
In this dissertation, two novel cross-coupling control (CCC) schemes derived from different approach are proposed to enhance the contouring accuracy of direct-drive robots (DDR). Considering that DDR abound in model uncertainties and external disturbances, the main purpose of the proposed CCC schemes is to achieve excellent contouring accuracy and guarantee stability simultaneously in the presence of the model uncertainties and external disturbances. Firstly, a CCC using scheme is proposed; from the relationship between the coupled and uncoupled system, the design of the compensator in the CCC scheme can be transformed onto an equivalent feedback control design problem which has time-varying (TV) parameters in. scheme is adopted to design the cross-coupling (CC) controller to provide excellent contouring accuracy and achieve internal stability even when the CC gains and the model parameters are varying. Secondly, instead of directly calculating the instantaneous contouring error and feeding it to coordinate axes, a novel Lyapunov-based CCC scheme is proposed to enhance contouring error by reducing the error normal to the contour. Based on Lyapunov backstepping method and the recursive updating technique, the proposed control scheme can enhance the contouring accuracy for three-dimensional contouring in spite of considerable model uncertainties, disturbances, and even gravitational effects. Furthermore, the proposed CCC design, which is a typical MIMO nonlinear TV system, can be verified as locally stable. Lastly, a two-link and a three-link anthropomorphic ultrasonic motor (USM) serial DDR systems are established to investigate the feasibility. Experimental results validated the efficacy of the proposed CCC schemes and the results showed that the proposed CCC schemes could maintain excellent contouring accuracy under various operating conditions such as contouring commands, payload configurations, and robotic configurations.