Target detection systems using multiple sensors are typically required to process large amounts of data. However, significant data compression is possible if the relatively new technique of decentralized detection is used. In this technique, decisions made at the sensors level are transmitted to a central processor via communication links for decision fusion. System performance then becomes strongly dependent on reliable transmission through the decision channels. In this proposal, therefore, we address the optimal design of communication links and the development and analysis of communication protocols with feedback that permit reductions in average data rates on the decision links . This will result in optimized system performances, smaller average transmitted power requirements for the sensors, and a measure of immunity to eavesdrop and intelligent jamming. For each of the above two problems, we intend to perform relevant theoretical research followed by development and execution of computer programs that calculate overall system designs and associated performances. We shall present ample numerical results in support of theoretical predictions.
