Investigation Of Iterative Learning Control Theory Algorithm

Iterative learning control (ILC for short) theory is an important branch of intelligence control field. Its fundamental principle is to acquire the current iterative input based on the former input signals and the tracking error's proportional, integral and derivative components. ILC can realize the tracking of a setting output trajectory within a prescribed accuracy during the given period of time after several iterations.This paper's main work includes the following aspects:1. Put forward high order open and closed loop ILC to solve a kind of linear time-varying systems and nonlinear systems. This arithmetic not only utilizes the current iteration tracking error, but also uses several preceding tracking error signals. By using more information it can improve the learning abilities of the being studied systems. This method has a faster convergence speed than the pure high order one. We present the convergence conditions and prove this arithmetic is convergent under these conditions. The effectiveness of the proposed ILC scheme is illustrated by a simulation.2. Owing to the tracking error bound is related to the initialization error bound an initial state learning scheme is proposed together with the high order open and closed loop ILC updating law in this paper. This scheme can eliminate the effect of the initialization error on the convergence of ILC gradually. The simulation results indicate that this scheme may give an improved ILC transient performance alone the ILC iteration number direction.3. We analyze a kind of discrete time linear time-invariant system with impulse response and design its iterative learning controller PID optimal parameter based on certain performance index. These parameters are gained according to the specialty of the plant to be controlled. So these parameters can be calculated by mathematic method unlike the parameters selection based on a trial-and-error method in the former ILC algorithm.4. We extend this parameter optimal design method to close loop control. As the closed loop control being added, we accelerate the ILC convergence speed. Otherwise the ILC convergence speed is rather slow for unstable systems by only using open loop parameteroptimal design. This open and closed loop algorithm has better performance in convergence as for unstable systems which can be seen from the results of the simulation experiments5. Several simulations have been made according to the ILC algorithms that been proposed in this paper. These simulation results illustrate the feasibility and effectiveness of the proposed iterative learning controller.