摘要
针对一类迭代学习控制提出了一种基于二次性能指标函数的自适应参数优化方法。如果原离散系统是正定的,那么这种具有可调参数的学习算法可以保证误差按几何单调收敛于0,如果系统非正定的,提出了一种反馈调节方法使系统正定。数值仿真表明了所提出算法和条件的有效性。
A class of iterative learning control with adaptive parameter optimization is proposed through a quadratic performance index. This learning algorithm with adjustable parameter guarantees geometrically monotonic convergence of the error to zero if the original discrete-time system is positive. If it is not positive, the conditioning method with feedback makes the system positive. Numerical simulations are presented to illustrate the effectiveness of the algorithm and condition.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2005年第8期1833-1835,共3页
Journal of System Simulation
关键词
迭代学习控制
参数优化
单调收敛
正定
iterative learning control
parameter optimization
monotonic convergence
positive
