Input-to-State Stability of Packetized Predictive Control Over Unreliable Networks Affected by Packet-Dropouts

This work studies a predictive networked control scheme in which packets containing optimizing sequences of control inputs are sent over an unreliable communication network affected by data-loss. We show that, provided the number of consecutive packet losses is bounded, input-to-state stability can be ensured by appropriate choice of design parameters. Our results apply to the general case of constrained nonlinear plants which are affected by uncertain and unmeasured disturbances.