Dynamic feedback control over unreliable communication channels

This paper investigates the problem of dynamic feedback control over unreliable communication channels with non-stationary packet loss. The rate of control and observation signals are modelled by two mutually independent random variables. In particular, the problem is considered for two control packet rate compensation schemes. In one scheme, called as a zero-input scheme, where the control input packet applied by the actuator sets to zero when the control packet dropped out. In the other scheme, called as a hold-input scheme, the previous control input is applied by the actuator when the control packet dropped out. Classes of observer-based stabilizing controllers for networked control system are designed with the observer and the feedback control gains are being determined by solving linear matrix inequalities. Numerical simulations are performed to confirm the effectiveness of the proposed design procedure.