Distributed Bayesian hypothesis testing with distributed data fusion

The problem of distributed Bayesian hypothesis testing with distributed data fusion is examined. In the distributed data fusion configuration, some signal processing is done locally at the sensors and the partial results are transmitted to the other sensors for further processing and fusion. Global results are obtained at each of the sensors. This system configuration is attractive for many applications from the survivability point of view. The problem is formulated, and optimum decision rules and fusion schemes are obtained that minimize the Bayesian risk at each sensor. An example is presented for illustration.>