Specified QoS based networked observer and PI controller design with disturbance and noise rejection under random packet dropout

Abstract

In this paper, a novel observer based networked PI controller has been designed by considering bounded disturbance and measurement noise under specified probability of random packet dropouts in networked control systems (NCSs). Lyapunov stability condition of the NCS with an observer and PI controller has been derived using an asynchronous dynamical system (ADS) approach as a linear matrix inequality (LMI) with nonconvex constraints to improve its quality-of-service (QoS) and quality-of-control (QoC). In order to improve QoC and QoS for NCS under process and measurement disturbance, we have used H∞ norm bound, and random packet drop rate in the NCS design. The algorithm has been tested and validated on four test-bench models of energy conversion systems. Confidence intervals and the mean responses of the state variables, state estimation errors of the observer and control signals have been reported using 100 Monte Carlo simulations. Our proposed NCS with PI controller and observer design method can be employed on a wide range of state space models of energy conversion systems. The specified probability of packet dropouts in the communication network between the sensor, observer, PI controller and actuator has been considered in the design which allows design of the NCS with guaranteed Lyapunov stability. We consider a switched linear dynamical system structure of the NCS with bounded noise and disturbance inputs while jointly searching for PI controller and observer gains. Using the controller and observer designed with 90% packet drop rate, we show the system performance degradation at 97%–99% QoS using phase portraits of the error dynamical system around the equilibrium as well as through univariate and multivariate hypothesis tests for equal dispersion around the equilibrium of the error dynamical system.