Event triggered estimator based controller design for networked control system

This paper presents an event-triggered estimator-based feedback control design for networked control systems (NCS) under the inevitable stochastic effects and correlated noises. Stochastic effects are modelled using random variables that obey Bernoulli distribution. There always exists a trade-off between control performance and usage of the resource. Thus, at the sensor node the relative errorbased event-triggered mechanism (ETM) is incorporated to regulate the data transmission, thereby reducing energy consumption and network traffic. Here, a predicted sensor measurement is used by an estimator if data packet is lost and the event failed to trigger. This will improve the performance and reduces the computation burden. The proposed estimator estimates the state information using orthogonal projections that completes feedback control design. Sufficient conditions are provided for the mean-square stability of the proposed method using Lyapunov theory. Subsequently, the proposed method's steady-state property, energy consumption, and computation burden are discussed in detail. Hence, it is worthy to note that the proposed work is suitable for online estimations because of its recursive structure. It is also suitable for a system with model mismatch and correlated noises. The efficacy of the proposed work is demonstrated using real-world applications in the simulation platform. (c) 2021 ISA. Published by Elsevier Ltd. All rights reserved.

Automated summary

This paper presents a feedback control design based on an event-triggered estimator for networked control systems (NCS) under stochastic effects and correlated noises. The proposed method incorporates an event-triggered mechanism at the sensor node to regulate data transmission and reduce energy consumption and network traffic. The estimator uses a predicted sensor measurement in case of data packet loss, improving performance and reducing computation burden. The proposed method achieves mean-square stability and is suitable for online estimations with model mismatch and correlated noises.