Optimization of the distributed detection system with data fusion

The optimal decision fusion with dissimilar local detectors is analyzed. Applying the central limit theorem, it is demonstrated that the asymptotic probability of false alarm is zero and the asymptotic probability of detection is one. In addition to the analysis on optimal data fusion rule, a computationally efficient solution for finding the optimal set of local thresholds is proposed. It is based on the concept of rate distortion. The optimal solution gives the minimum mean square of quantization error. This results in solving a set of O(N ) nonlinear, coupled equations, where N is the total number of local detectors. In comparison with previous works, which require solving a set of O(2^N) nonlinear equations, the authors´ approach is expected to lead to savings in terms of computational time and complexity