Mode dependent H∞ predictive control of networked control systems under variable-period sampling

An H∞ predictive control problem is investigated for a class of networked control systems with random delays and disturbances. Applying the variable-period sampling method, the networked control system is discretized and modeled as a Markovian jump linear system with partly unknown transition probabilities. By using the one-step cost function, sufficient conditions are presented to guarantee asymptotical stability and an upper bound of the performance index for the closed-loop system. The design procedure for the state feedback controller is also derived based on the linear matrix inequality method. Finally, an example is presented to illustrate the effectiveness of the proposed method.