Model-based event-triggered control for linear systems using a finite-time convergent observer

Abstract

In this paper, we present an improved model-based event-triggered control approach for regulation of networked linear time-invariant plants by integrating a finite-time convergent observer. Utilizing the send-on-delta concept, we introduce two event-triggering strategies that substantially decrease the number of samples transmitted over the sensor-to-controller communication channel: (1) one strategy guarantees global asymptotic stability of a tunable compact neighborhood of the origin with a finite number of sampling events; (2) the other strategy ensures global asymptotic stability of the origin with arbitrarily large inter-sampling time. Through the lens of hybrid systems theory, we demonstrate that both methods avoid Zeno solutions, making them suitable for digital implementation, and both strategies exhibit robustness against both small and large perturbations. We showcase the advantages of our proposed strategies over other model-based approaches using a perturbed linearized batch reactor example.