Kaiman filtering with scheduled measurements — Part II: Stability and performance analysis

For the purpose of reducing communication rate, we have proposed two types of measurement innovation based scheduling algorithm for state estimation of linear discrete-time stochastic systems in our companion paper. More specifically, the first one is an iterative scheduling algorithm, which sequentially triggers sensor communication during each sampling interval. The second one is simpler and only the vector measurement outside a deadzone will be communicated to estimator. This paper is devoted to stability and performance analysis for the derived optimal estimator under the above schedulers. Necessary and sufficient conditions for stability of the established estimation framework are established. Moreover, it is shown that under a given communication rate, the first scheduling algorithm outperforms the second one. An illustrative example is included to validate our theoretic results.