非线性系统开闭环D型迭代学习算法的鲁棒性

研究了系统满足收敛条件时，具有开闭环D型迭代学习控制律的一类非线性系统在动态过程存在干扰的情况下控制算法的鲁棒性问题。理论分析表明，当系统动态过程扰动有界时，开闭环D型迭代学习控制算法是鲁棒的。当满足开闭环学习收敛条件时，控制误差收敛到期望值的某一个邻域，其大小与相邻两次迭代运行中这些干扰的大小有关，而与初始控制输入等无关，当干扰越小时，学习控制过程越接近于期望值。

ROBUSTNESS OF OPEN-CLOSED-LOOP D-TYPE ITERATIVE LEARNING CONTROL ALGORITHM FOR NONLINEAR SYSTEMS

To analyses robustness of open-closed-loop D-type iterative learning control algorithm for a class of non-linear system in the paper, when the system satisfy the convergence condition of system, and there exists disturbance in the dynamic process. The theory analysis indicates, when disturbances of the dynarnic process were bounded. ILC scheme is robust. The control error will converge to some neighborhood of desired control input, which value is associated with the disturbance of consecutive two iterative circles, when the scheme satisfies convergence condition. The neighborhood is independent of the initial control input. The less the disturbance, the closer the control input would approach desired control input.