Performance of search via track-before-detect for distributed sensor networks

We consider the search for moving targets over large areas using a network of fixed sensor nodes. The concept of track-before-detect is defined and used to both manage the information between sensors and reduce the likelihood of false searches. We develop expressions for the probability of search success and the probability of reporting false search in this system concept. Using these as performance measures, we examine how the track-before-detect search strategy impacts design choices in these networks, by showing which parameter changes in the sensor design have the greatest impact on improving the desired performance goals.