Distributed multiple hypotheses testing with serial distributed decision fusion

In this paper we present a multiple hypothesis testing strategy for serial (so-called tandem) distributed decision networks. Comparing with parallel decision networks, some nice properties of tandem networks are also of interest. Applying our flexible multiple Hypotheses Testing methods to serial distributed multisensor systems, the decision risk functions are provided, the decision rules for each individual sensor (viewed as middle nodes) as well as fusion sensor (as final node) are derived. It is also shown how to compute sensor performance and system performance in terms of variety of probability masses. A numerical example is given to demonstrate the efficacy of the proposed method.