Research On Trajectory Tracking Based On Adaptive Iterative Learning Control

Adaptive iterative learning control is an important research direction of intelligent control.Adaptive iterative learning control has good control performance for the uncertain system model,which is of great significance to the productive practice.According to the previous control experience and the error of system output signal and desired output signal,an ideal control input signal can be obtained.After several iterative controls,the controlled objects can track the given trajectory with high precision in given time.Adaptive iterative learning control only needs few prior knowledge and have great tracking effects.What's more,most of the controlled current systems are nonlinear,strong coupling and difficult to be modeled,such as robotic manipulators.Therefore,the research of adaptive iterative learning control is of great significance.This article mainly focuses on the random of iterative initial point of the adaptive iterative learning control system,the incomplete repetition of the iterative locus and the non-consistent of the iterative length.Firstly,this paper introduces the application of adaptive iterative learning control method in the robotic manipulator system.Then the problem of the repetitive trajectory of uncertain manipulator parameters is analyzed.The controllers are designed for the overlapping trajectories of the initial state,the space closed trajectory and the random initial state point,respectively.The analysis demonstrates that the proposed method can achieve perfect tracking trajectory that even when the initial trajectory error is not zero.For the incomplete repetition of the iterative locus of the adaptive iterative learning control system,the function factor is introduced to the initial error to suppress the initial error effects on the system performance.The tracking error of the system converges to zero with the increasing of iterations.All the variables of the system are uniformly bounded in the process of running.In addition,Lyapunov-like analysis method to design the iterative learning controller is adopted in the paper to avoid the basic conditions of Lipschitz.For the non-consistent of the iterative length of the adaptive iterative learning control system,an equivalent error is used to deal with the different iterative time lengths of different tracking tasks,which effectively avoid the requirement of the classical AILC that the iteration length must be the same.Finally,the proposed manipulator model is simulated.Both the theoretical and simulation results demonstrate the effectiveness of the proposed algorithm.