Performance evaluation of decentralized estimation systems with uncertain communication

An approach for evaluating the performance of decentralized estimation systems under non-ideal communications is presented. Recent studies have shown that transmission disruptions occur frequently in tactical wireless networks due to a variety of unpredictable factors, such as unintentional interference or jamming. Yet the impact of non-ideal communications on algorithmic performance are often ignored in the fusion and tracking literature. This paper provides the beginnings of a basis for robustly assessing decentralized estimation system performance in the presence of uncertain communications through the development of a context metric - the expected information rate - that characterizes problem difficulty in terms of the communication environment.