Distributed detection by a large team of sensors in tandem

The problem of decentralized binary hypothesis testing by a team consisting of N decision makers (DMs) in tandem is considered. Each DM receives an observation and transmits a binary message to its successor; the last DM has to decide which hypothesis is true. Necessary and sufficient conditions are derived for the probability of error to asymptotically (as N→∞) go to zero. The result is generalized for multiple hypothesis and multiple messages. An easily implementable suboptimal decision scheme is also considered; necessary and sufficient conditions for the probability of error to asymptotically go to zero are derived for this case as well. The tradeoff between the complexity of the decision rules and their performance is examined, and numerical results are presented