Output-feedback control with wireless channel state detection and actuation message dropout compensation

This paper presents a framework for designing optimal output-feedback controllers that use wireless sensing and actuation links with imperfect channel-state information. Remote system state estimation is performed using a prediction–correction filter that resembles the traditional Kalman filter and incorporates current measurement inputs. The controller computes the current and tentative future control inputs based on the estimated remote system state and the detected wireless channel state. These control inputs are transmitted to actuators as messages. The message dropout compensation strategy for actuation involves scaling the most recent control input when no previously received tentative control inputs are available. We analytically solve finite- and infinite-horizon output-feedback control problems and prove the validity of the separation principle, assuming a reliable mechanism for acknowledging actuation message transmission. We validate the results using an illustrative numerical example that demonstrates the practicality and effectiveness of our framework.