Kalman filtering with scheduled measurements - part I : estimation framework

Abstract

This paper proposes an estimation framework under scheduled measurements for linear discrete-time stochastic systems. Both controllable and uncontrollable schedulers are considered. Under a controllable scheduler, only the normalized measurement innovation greater than a threshold will be communicated to the estimator. While under an uncontrollable scheduler, the time duration between consecutive sensor communications is triggered by an independent and identically distributed process. For both types of scheduler, recursive estimators that achieve the minimum mean square estimation error are derived, respectively. Moreover, necessary and sufficient conditions for stability of the mean square estimation error are provided.