Iterative learning control for discrete-time switched systems with attenuation factor

Abstract

In order to decrease the effects of measurement noise on the trajectory tracking control of discrete-time switched systems, this paper proposes a discrete iterative learning control algorithm with an attenuation factor. The proposed algorithm adds a learning gain attenuated along the iteration horizon into measurement errors interfered by measurement noise for modifying the control rules of switched systems, in order to decay measurement noise as iterations increase. The convergence of each subsystem is proven rigidly with λ-norm theory, and the convergent condition of switched systems is provided. Theoretical results indicate that the proposed algorithm can effectively suppress non-repetitive measurement noise, and realize the complete tracking of the desired trajectory for the output of a discrete-time switched system within limited time. The final simulation results verify the validity of the proposed algorithm.

Iterative learning control for non‐linear switched discrete‐time systems

“…Pre-and post-multiplying both sides of (47) by diag{P i , Q i , P i , P i , Q i } and substituting (11) into (47) yield (37). It comes from (38), (41) and (42) that the condition (27) holds. Thus by Theorem 2, we can conclude that the system (10) is exponentially stable under the switching signal σ(t).…”

“…A highorder ILC scheme for non-linear systems with parametric uncertainties was proposed in [26]. To decrease the effects of measurement noise on the trajectory tracking control of discretetime switched systems, an ILC algorithm with an attenuation factor was proposed in [27]. Besides, ILC method was also applied to the switched systems [12,28].…”

Learning Gain Self-Regulation Iterative Learning Control for Suppressing Singular System Measurement Noise

“…However, the method in [23] uses integral and fractional derivative, which makes the design of iterative learning controller complicated. In order to reduce the effect of measurement noise on the tracking accuracy of the system, the iterative learning control algorithm based on the attenuation of the learning gain along the iteration axis is proposed in [24], [25], which effectively suppresses the non-repeated measurement noise. Although the two algorithms retain the advantage of simpleness of the iterative learning controller, none of them can adjust the size of the learning gain based on the tracking error.…”

Iterative learning control for linear continuous-time switched systems with observation noise

“…For instance, some ILC strategies have been proposed by Bu et al (2013a, 2013b, 2013c), Yang (2015) and Cao and Sum (2014). Furthermore, an improved ILC scheme with attenuation factor has been addressed by Cao and Sum (2014). Reviewing the above studies, we found that most focus on analysing convergence or stability by employing the λ -norm theory and the lifted-system theory.…”