论文部分内容阅读
针对存在扰动的未知非线性系统,利用小波逼近将系统参数化,结合变结构控制技术,提出了一种鲁棒迭代学习控制算法.该算法采用迭代学习的方式修正小波逼近的系数,利用具有死区的滑模变结构技术保证算法的鲁棒收敛性.收敛性分析表明,每次迭代学习都将减小所得到的逼近系数与最佳系数的差异.因此,期望轨迹变化后,该算法针对以前轨迹的学习结果仍然可以起作用,部分克服了传统迭代学习控制的学习结果仅对某一特定轨迹有效的缺点.
Aiming at the unknown nonlinear system with disturbance, the system is parameterized by using wavelet approximation, and combined with variable structure control technology, a robust iterative learning control algorithm is proposed, which uses iterative learning method to modify the coefficient of wavelet approximation, The sliding mode variable structure technique in the region ensures the robust convergence of the algorithm. The convergence analysis shows that each iterative learning will reduce the difference between the obtained approximation coefficient and the best coefficient. Therefore, after the expected trajectory changes, Previous trajectory learning results are still valid, partly overcome the shortcomings of the traditional iterative learning control learning results valid only for a particular trajectory.