ABSTRACT
In this chapter, we consider more general multisensor multihypothesis decision systems such as tandem, hybrid/tree network system, and the networked systems with a given transmission error probability, i.e., with a fading channel. The results on optimal sensor rules and the unified fusion rule for parallel network systems with the ideal channel can be extended to these general network systems. Besides, another popular decision fusion strategy—optimal decision fusion with given sensor rules is presented. Then, by combining the results of the two types of decision formulations, for the general systems where the fusion center is not necessarily a observer, we propose an efficient simultaneous searching algorithm for both optimal sensor rules and a unified fusion rule in Section 3.5.1. Numerical examples support these results and give some guidance as to which sensor one should choose as the fusion center to achieve better performance. Finally, we present a counterintuitive result on performance analysis of communication direction in the last sections which is of some significance to design the communication direction of a networked decision system.
