Model-Based Event-Triggered Control over Lossy Networks

TL;DR

This paper proposes a novel event-triggered control method for networked systems over lossy networks that does not rely on acknowledgement messages, ensuring stability and positive inter-event times despite packet dropouts.

Contribution

It introduces a new approach for event-triggered control that relaxes the need for acknowledgements and handles packet losses in uncertain systems.

Findings

System can be asymptotically stabilized despite packet dropouts.

A positive minimum inter-event time is guaranteed.

The method does not require acknowledgement messages.

Abstract

The event-triggered control problem over lossy communication networks is addressed in this paper. Although packet dropouts have been considered in the implementation of event-triggered controllers, the assumption of protocols that employ acknowledgement messages persists. This paper provides an approach that relaxes such assumption. An event-based controller is implemented at the sensor node of the networked and uncertain system and it transmits feedback measurements to the controller node at asynchronous time instants. The transmitted packets of information are subject to dropouts by the lossy network. We show that the uncertain system can be asymptotically stabilized, that a positive minimum inter-event time exists, and that the proposed approach does not require acknowledgement messages.