目标轨迹更新的点到点鲁棒迭代学习控制

Point-to-point robust iterative learning control via reference trajectory regulating

为了获得更快的算法收敛速度,提出易于操作的插值法来更新目标轨迹.该方法易于选择参数,有更好的泛化性能并为控制算法的设计提供灵活性.采用有更广泛收敛域、能够更好地抵消噪声影响的P型闭环迭代学习控制.对于有初始状态扰动的情况,新算法结合了P型闭环迭代学习控制与插值法轨迹更新,收敛性和鲁棒性得到理论证明.数值仿真验证了新算法比固定目标轨迹算法和P型开环迭代学习控制更有优势.

An interpolating algorithm was proposed to regulate reference trajectory in order to achieve faster convergence rate.The algorithm has the character that choosing parameter more intuitively and not relating to specific ILC algorithm.P-type current cycle ILC was utilized because the predominant performance such as more wide convergence interval and better robustness.The new algorithm combines current cycle ILC and reference trajectory regulating to absorb both of their advantages.The convergence and robustness of the algorithm were rigorously analyzed.Numerical simulations verified the new algorithms superiority over fixed reference trajectory and open-loop ILC.