Servo Motor Current Feedback Control to Optimize the System including Reduction Gear Dynamics by Using Adaptive Control Technique

点击翻译按钮获取中文摘要

In this project, several schemes to accomplish high performances of speed and accuracy of DC servo motor control are given. These schemes are used in an adaptive control method. The adaptive method is, first, to identify dymanics of reduction gear, then, adapt the currenct control feedback gain to the dynamics and compensate disturbances such as friction forces optimally.Following schemes of identification, estimation and control for the adaptive method are proposed.1. Two identification schemes are proposed. Those are discrete-time and continuous-time identification law based on Lyapunov-type adaptive law, which is computed in a short time period. To improve convergence property, a non-linear term is added to the discrete-time law and relay elements are used in the continuous-time law.2. A sufficient condition for control systems to be stable is given. Applying this condition, it is shown that the systems icnluding the proposed identification schemes are stable.3. For servo systems with nonlinear dynamics an identification scheme is proposed. By using filter functions, the schem does not need to use acceleration signals. The scheme is tested by identifying parameters of an experimental manipulator with 6 links.4. An observer estimating periodical disturbances such as friction forces in reduction gears is given. The observer includes an identification law to identify the frequency of the disturbances.5. It is shown that the adaptive control system using the proposed identification law with relay elements is robust to disturbances such as friction forces. The fact is confirmed by on-line control experiments using the experimental manipulator.6. A decoupling adaptive control method is given. The method compensates nonlinear terms in the servo systems by using estimated paramewters and reduces control erroes caused by estimation errors by disturbance decoupling method. Experimental results using the manipulator show reduction of control errors.