H2 Optimization Under Intermittent Sampling and its Application to Event-Triggered Control

Abstract The paper studies the H2 optimal sampled-data control of linear time-invariant systems. The sampling pattern is assumed to be uniformly bounded but otherwise arbitrary and a priori unknown. The derived optimal solution is analytic, computationally simple, implementable, and transparent. Properties of the optimal controller also suggest novel event triggering mechanisms, which preserve stability and tolerant of measurement noise. Applying to the benchmark problem of Cervin (2016), the proposed event generation mechanisms outperform several existing approaches.