Kalman Filter for Networked Control Systems Synchronized at Actuation Instants

Traffic and latency in the communications network, as well as processing of the control algorithms cause sampling-actuation variable delays (SAVDs) in networked control systems (NCSs). Moreover, feedback control loops implemented via NCSs, may suffer degradation in control performance and even become unstable due to SAVDs. To remove these negative effects on per-formance/stability, a method based on strictly periodic actuations and aperiodic samplings has been featured in the literature. In this approach, the states are measured at the sampling instant, the values that these states will have at the actuation instant are then estimated, and subsequently, the control action is calculated based on these estimates. This paper extends the above model to the case of noisy samples and disturbed process by coupling a Kalman filter. Additionally, it is also solved the problem of deciding instants and periods in which the predict and update phases of the Kalman filter are applied. It has been obtained that the synchronized actuations eliminate the harmful effects of SAVDs in NCSs. Moreover, the Kalman filter improves the performance under noisy measurements and process disturbance.