分数线性相位超前迭代学习控制  

Fractional linear phase-lead iterative learning control

作　　者：潘雪[1] 叶永强[1] 王建宏[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016

出　　处：《控制理论与应用》2013年第7期869-874,共6页Control Theory & Applications

基　　金：国家自然科学基金资助项目(61074161;61034005);江苏省自然科学基金资助项目(BK2010507);教育部博士点基金新教师类资助项目(20103218120014);江苏省普通高校研究生科研创新计划资助项目(CXZZ12 0158);南京航空航天大学博士学位论文创新与创优基金资助项目(BCXJ13 07)

摘　　要：线性相位超前迭代学习控制的超前拍数需要取整,不利于控制系统的性能达到最优.对此提出分数线性相位超前迭代学习控制.对系统的收敛条件进行频域分析,得到超前拍数,学习增益和可学习带宽的范围.通过调节超前拍数来提高系统的可学习带宽,以达到更高的跟踪精度.给出了分数线性相位超前的具体实现方法,并在此基础上对整数与分数相位超前的补偿效果进行比较.以机械臂为被控对象的仿真结果表明了分数线性相位超前更能提高系统的可学习带宽及跟踪精度.Because the number of lead steps in a linear phase-lead iterative learning control must be an integer which hampers the optimization of a control system,we propose a fractional linear phase-lead iterative learning control method.The range of the lead step,the learning gain,and the learnable band can be calculated by analyzing the convergence condition in the frequency domain.By adjusting the number of lead steps,we can achieve a wider learnable band and a higher tracking accuracy.The detailed realization of the fractional lead is introduced,and the comparison of the compensation effects between it and the integer lead is given.A comparison simulation in controlling a robot arm-joint is carried out.The results show that the fractional linear phase-lead iterative learning control provides wider learnable band and higher tracking accuracy.

关 键 词：迭代学习控制 分数线性相位超前 频域分析 超前拍数 跟踪 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]