引用本文:严求真,孙明轩,蔡建平.含输入死区系统的滤波误差初始修正迭代学习控制[J].控制理论与应用,2017,34(1):77~84.[点击复制]
YAN Qiu-zhen,SUN Ming-xuan,CAI Jian-ping.Filtering-error rectified iterative learning control for systems with input dead-zone[J].Control Theory and Technology,2017,34(1):77~84.[点击复制]
含输入死区系统的滤波误差初始修正迭代学习控制
Filtering-error rectified iterative learning control for systems with input dead-zone
摘要点击 2679  全文点击 2140  投稿时间：2016-03-24  修订日期：2016-12-27
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2017.60161
  2017,34(1):77-84
中文关键词  死区  迭代学习控制  初值问题  非参数不确定性  Lyapunov方法
英文关键词  dead-zone  iterative learning control  initial condition problem  nonparametric uncertainties  Lyapunov approach
基金项目  国家自然科学基金项目(61374103, 61573320, 61573322), 浙江省公益技术应用研究项目(2016C32093, 2017C33155), 浙江省教育厅科研项目 (Y201635861)资助.
作者单位E-mail
严求真 浙江水利水电学院信息工程学院 zjyqz@126.com 
孙明轩* 浙江工业大学信息工程学院 mxsun@zjut.edu.cn 
蔡建平 浙江水利水电学院应用数学研究所  
中文摘要
      针对一类输入环节含死区非线性特性且误差初值非零的非参数不确定系统, 提出滤波误差初始修正学习 控制方案, 分别解决死区斜率下限可知与未知两种情形下的轨迹跟踪问题. 给出了两种修正滤波误差信号构造方 法, 并根据Lyapunov综合方法设计学习控制器, 采用鲁棒学习策略处理非参数不确定性和死区非线性特性. 经过足 够多次迭代后, 实现滤波误差在预设的作业区间也收敛于零. 文中所提出的控制方案, 具有构造简单与实施方便的 特点, 仿真结果表明了本文所提控制方法的有效性.
英文摘要
      This paper presents the filtering-error rectified adaptive iterative learning control algorithms to tackle the trajectory??tracking problem for a class of nonparametric uncertain systems with unknown input dead-zone, in the presence of arbitrary initial states. To overcome the arbitrary initial states, two construction programs of the rectified filtering-error are proposed. Two iterative learning controllers are designed by applying Lyapunov synthesis, suitable to the case that the lower bound of the dead-zone’s slope is known and the case that it is unknown respectively, dealing with the nonparametric uncertainties and the unknown dead-zone nonlinearity according to the robust learning strategy. As iteration increases, the filtering error converges to zero on the specified interval. The rectified filtering-error signal can be simply constructed, and the proposed learning control scheme, whose effectiveness is demonstrated in the presented numerical results, is easy for implementation.