Analysis of a Two-Sensor Tandem Distributed Detection Network

摘要

The following distributed detection problem is formulated. We consider a team that comprises of two decision makers, who are referred to as Decision Maker A (DMA) and Decision Maker B (DMB). Both decision makers receive uncertain measurements or observations, and the goal of the team is to make a decision with the objective of trying to minimize the probability of making an incorrect decision. DMA processes his measurement first and communicates to DMB one of K messages, M1, M2,....,MK. Based on this message and his own observation, DMB makes the final decision of the team. The goal is to analyze the performance of the above scheme (using values of K greater than two) and compare it with the well known case where only two messages, M1 and M2, are used by DMA. It is interesting to see how the performance of the team approaches that of the centralized version of the problem (i.e., two independent observations available to a single decision maker) with increasing values of K. Furthermore, results for the General K case have been presented and can be used to evaluate the performance of a team where K messages (K taking on any value) are used for communication between DMA and DMB. (RH)