任意初态下的扰动系统学习控制

针对任意初态下带有扰动的线性定常系统, 提出了相应的控制算法。该算法将受控过程分成无穷个子过程, 系统利用上一子过程的输入输出信息来调节当前过程的输入，以期获得更好的控制效果。在控制过程中, 基于迭代学习控制思想, 借助初始修正手段, 可使系统的跟踪误差达到任意小, 且当系统无扰动时，在指定区间内可实现完全跟踪。最后, 通过仿真算例, 验证了算法的有效性。

Learning control for the disturbance system with arbitrary initial state error

A control algorithm for the linear time-invariant system is presented in this paper. It divides the controlled process into infinite subprocesses. For each subprocess, the system uses the input and output information of the last subprocess to adjust the current input in order to achieve a better control effect. In the control process, based on the iterative learning control theory, this algorithm solves the control problem using the initial rectifying methods. It can make the tracking error of the system converge to an arbitrarily small range, and achieve complete tracking if there is no disturbance in the specified time period. Finally, the effectiveness of the stated algorithm is demonstrated by the simulation result.