Multi-Sensor Scheduling for State Estimation with Event-Based, Stochastic Triggers*

In Networked Control Systems, remote state estimation is hampered by limitations in sensor to estimator communication. Past approaches involving scheduling sensors dynamically via a deterministic event-triggering mechanism reduce communication while maintaining estimation quality However, these approaches destroy the Gaussian property of the innovation process, making it computationally intractable to obtain an exact minimum mean squared error (MMSE) estimate. Recent work has proposed utilizing a stochastic event-triggered sensor schedule for state estimation. We extend these results to the multi-sensor case, obtaining closed-form expressions for the MMSE estimator and its covariance matrix as well as performance bounds for the system.