Abstract
Abstract
Robust monotonic convergence (RMC) should be obtained when designing ILC (Iterative learning control) system. But it may cause convergence rate becoming slower. In order to maintain stability of control system, transient growth is often limited within a certain range, which is one of the most properties of transient behaviors. Pseudospectral of the system transfer matrix is introduced to predict the transient growth of ILC system. At the same time, asymptotic performance can be improved if the transient growth is kept small. Under this case, the convergence rate will also be speeded up. The simulation demonstrates the effective of new method for designing ILC system.
Reference17 articles.
1. A Survey of Iterative Learning Control;Bristow;Control Systems IEEE,2006
2. The Phenomenon of Apparent Convergence Followed by Divergence in Learning and Repetitive Control;Longman;Intelligent Automation & Soft Computing,2002
3. Techniques for Improving Transients in Learning Control Systems;Chang,1992