摘要
针对一类具有不确定性的多输入多输出非线性系统,提出一种迭代学习控制算法.该算法具有的特点是:针对任意初态情形,结合开环D型迭代学习控制器的优点,在时间轴上设计了一个随迭代次数增加而缩短的时间段.在该时间段上,控制算法对状态偏差进行修正,以使系统输出在此段时间后跟踪期望输出,且系统跟踪误差收敛到一个界内.这个界仅由系统自身不确定性和不确定的外界干扰决定,与初态误差无关.当外界扰动为0,以及迭代次数趋于无穷时,经过上述时问段后,系统输出精确跟踪期望输出.理论证明和仿真结果都说明了该算法的有效性.
A novel iterative learning control technique is presented for a class of multi-input and multi-output nonlinear systems with uncertainty. It is well known that the system output cannot track the desired output accurately and immediately due to the initial state error. In this paper, a special time interval, with the characteristics of more iterations and less time interval, is designed. By combining an open-loop D-type iterative learning controller, the random state errors are rectified in the special time interval. Subsequently, system outputs track the reference trajectory. Uniform bounds for the final tracking errors are obtained, which are only dependent on the system uncertainties and disturbances but independent of the initial errors. After the special time interval, the system outputs will accurately track desired outputs with infinite iteration and without disturbances. Both theoretical proof and simulation results show the effectiveness of the proposed algorithm.
出处
《自动化学报》
EI
CSCD
北大核心
2014年第7期1295-1302,共8页
Acta Automatica Sinica
基金
国家自然科学基金(61127006
61325017)资助~~
关键词
迭代学习控制
初始状态误差
快速学习
收敛性
鲁棒性
Iterative learning control
initial state error
fast learning
convergence
robustness
作者简介
吕庆南开大学机器人与信息自动化研究所博士研究生．主要研究方向为迭代学习控制及其在原子力显微镜中的应用．E—mail：1vq@robot．nankai．edu．cn
方勇纯南开大学机器人与信息自动化研究所教授．主要研究方向为视觉伺服，微纳米控制系统，非线性控制．本文通信作者．E—mail：yfang@robot．nankai．edu．cn
任逍南开大学机器人与信息自动化研究所博士研究生．主要研究方向为非线性系统控制，基于原子力显微镜的微纳米控制系统．E-mail：renx@robot．nankai．edu．cn
