Finite-Horizon ${\mathscr {H}}_{\infty }$ Fault Estimation of Time-Varying Systems with Random Access Protocol

Abstract

In this chapter, the finite-horizon $H_{\infty }$fault estimation problem is addressed for time-varying networked systems subject to the scheduling effect of Random Access Protocol. A sequence of independent and identically distributed random variables is adopted to model the sensor nodes obtaining access to the communication network under the scheduling effect of Random Access Protocol. The zero-input strategy is employed to compensate the data corresponding to the network nodes that have not been selected to transmit data. The faults considered in this chapter are supposed to be either incipient faults or abrupt faults, which are the most common faults in industrial processes. A finite-horizon $H_{\infty }$fault estimator is constructed to generate the estimate of states as well as the faults. Sufficient conditions are provided to ensure the prescribed fault estimation performance requirement. The desired estimator gains are characterized by means of the solutions of a family of recursive matrix inequalities.