Sampling multidimensional Wiener processes

摘要

In this paper, we formulate an optimal scheduling problem in continuous time for transmitting samples of a multidimensional Wiener process when there is a constraint on the communication between the observer and the estimator. We provide the optimal policy for both finite horizon and infinite horizon, and show that event-triggered sampling with constant threshold is optimal over infinite horizon when there is a hard constraint on the average frequency of communication.