摘要
针对在有限时间区间上重复运行的一类状态时滞非线性系统,研究了一种带有初态学习的迭代学习控制算法。基于算子谱理论证明了系统在任意初态条件下经过迭代学习后,其输出能够完全跟踪期望轨迹,得到了该算法谱半径形式的收敛条件,比范数形式的收敛条件放宽了。并通过对系统初态的迭代学习,放松了迭代学习控制算法对初始状态函数的要求。仿真结果验证了所提算法的有效性。最后,将该算法应用到含有状态时滞的非线性间歇过程,结果也表明了该算法的有效性和可靠性。
Aiming at a class of nonlinear system with state time-delay repeatedly running in limited interval, an itera- rive learning control algorithm with initial state study is studied. Based on operator spectrum theory, it is proved that the system output can track expected trajectory completely under arbitrary initial condition after iterations. Compared with norm form convergent condition, this paper relaxes the convergent condition of spectral radius. The algorithm relaxes the requirement for iterative initial state function after iterative learning. Simulation results indicate the validi- ty of the algorithm. Finally, the algorithm is applied in nonlinear batch processes with state delay. Results illustrate the effectiveness and reliability of the algorithm.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2012年第2期315-320,共6页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(61100103)资助项目
关键词
非线性时滞系统
迭代学习控制
初态学习
算子谱理论
nonlinear time-delay system
iterative learning control
initial state learning
operator spectral theory
