模糊增益PD型迭代学习算法及其应用

Fuzzy Gain PD Type Iterative Learning Control Algorithm and Its Application

普通PID型迭代学习控制算法由于其增益矩阵一旦确定,就不再改变,因而收敛速度较慢。为了提高收敛速度,结合模糊控制技术用于学习控制,提出一种模糊增益PD型迭代学习控制算法。该算法根据系统误差的大小,通过调整因子而改变增益矩阵大小。调整的方法是:在控制的初始阶段,增强控制输入中系统误差增益矩阵,同时保持误差微分增益矩阵不变,从而清除误差;而在控制趋于稳定时,增强控制输入中误差微分增益矩阵,同时保持系统误差部分增益矩阵不变,以减少超调量。针对一个单臂机械手模型,进行了仿真设计,仿真结果验证了方法的有效性。

The common Proportional Integration Differential （PID） type Iterative Learning Control （ILC） algorithm has a lower convergence rate since its gain matrix is constant. To improve the convergence rate, the fuzzy control law is introduced in ILC and a kind of fuzzy gain PD type iterative learning control algorithm is developed. The algorithm can regulate the gain matrices by adjusting the factors according to system error information. The adjustment method is ： 1 ） At the initial stage of control, enhancing the system error gain matrix while keeping the error differential gain matrix invariant, so the errors can be eliminated; 2） When control tends stable, enhancing the error differential gain matrix while keeping the system error gain matrix invariant, thus to reduce the overshoot. The simulation results for a single ann robotic model show that the method is feasible.