Reliable Distributed Estimation with Intermittent Communications

We consider the problem of distributed target tracking via networked sensors. Our setup consists of a set of sensors connected to a fusion center by means of communication links. Unreliable communication channels leads to communication delays and loss of information. To address this problem we model the arrival of messages from the sensors to the fusion center by a random process. We develop encoding/decoding rules for optimal decentralized estimation and present network topologies where centralized performance can be achieved. The main difficulty with optimal asynchronous decentralized tracking arises from both the process dynamics and the random arrivals. We show if packet arrival times are globally revealed the fusion center can realize optimal centralized performance but at a large computational cost. On the other hand if centralized performance is desired only at certain stopping times it can be achieved with scalable communication cost