The Sampled-Data Characteristic Model And Discrete Variable Structure Repetitive Control With Its Implementation

The sliding mode resulted by variable structure control performs well with strong adaptability when subjected to parameter perturbation and external disturbances. The tracking error can only converge to a neighborhood of the equilibrium point which is called the "quasi-sliding mode". Owing to the uncertain disturbance, perfect performance could hardly achieve by variable structure control, while repetitive control would be a good choice for the trajectory tracking in the presence of periodic disturbances. Performance improvement can be achieved when combining the variable structure control with repetitive control.This thesis presents the sampled-data characteristic model for the high-order dynamic systems, and the adaptive iterative learning control method is applied for the purpose of validation. By applying the ideal switching-dynamics strategy, the issues of variable structure repetitive control of discrete time uncertain systems are addressed, and the analysis of the closed-loop stability and convergence. The experimental results are presented to verify the effectiveness of the proposed methods.The main work and contributions in this thesis are as follows:1. According to sampled-data characteristic model theory, a second order time-varying difference equation is used to depict the systems. Least squares learning and iterative projection algorithms are introduced to identify the characteristic parameters.2. A nevol piecewise reaching law is suggested to form the ideal switching dynamics of systems. The variable structure controller designed can improve the chattering.3. For the reference trajectory which is periodic symmetry, the designed discrete variable structure repetitive controller can eliminate the periodic interference. In order to establish stability and convergence of the closed-loop systems, the monotone convergence layer, the absolute convergence layer and the quasi-sliding mode are discussed.4. The mathematical model of the linear motor is obtained by applying the least square algorithms. Based on it, the adaptive iterative learning controller and variable repetitive controller are designed. The experimental results of the sevo systems show the effectiveness of the strategies while the convergence performance of the closed-loop systems is guaranteed.