H{sub}2 FILTERING FOR DISCRETE-TIME PIECEWISE LINEAR SYSTEMS

摘要

This paper addresses the problem of H{sub}2 filtering for piecewise linear discrete-time systems. Both the system and filter have structures that change according to switching rules. Two types of switching rules are considered: (i) the system and filter switching rules are the same and based on an auxiliary signal which is supposed to be available on-line; (ii) the filter and system switching rules are different and based on the filter and system states, respectively. Based on piecewise quadratic Lyapunov functions, LMI conditions are derived for designing piecewise linear filters that ensure an optimized average estimation error variance.