Optimum Fusion in Distributed Multisensor Network Decision Systems

Abstract In this paper, we consider optional fusion rule in distributed multi-hypothesis multi-sensor network decision systems, where communication pattern between sensors are fixed and fusion center can observe data. We propose a specific fusion rule, which is of the most general form including all possible fusion rules as its a special case and independent of the statistical characteristics of observations and decision criterions; therefore, to reach the globally optimal decision performance, people only need to optimize sensor rules under this fusion rule for given conditional distributions of observations and decision criterion. Hence, we can regard this fusion rule as optimal fusion rule. Computer simulations support the above results.