Distributed Detection and Decision Fusion with Applications to Wireless Sensor Networks

This chapter focuses on distributed detection and decision fusion problems, which involve the design of decision rules at the local sensors and at the fusion center to optimize detection performance, under either a Neyman‐Pearson or a Bayesian criterion. It introduces the fundamentals of detection theory, and covers the conventional distributed detection problem with multiple sensors. The chapter presents several important factors that affect the design of distributed detection algorithms. They are the topology of the sensor networks, relation between sensor observations (conditionally independent versus correlated), optimization criteria, and quantization levels. Due to their high flexibility, enhanced surveillance coverage, robustness, mobility, and cost effectiveness, wireless sensor network (WSN) have great potential in environmental monitoring, battlefield surveillance, structural health management, among others. The described method is particularly useful when the marginal densities of sensor observations are non‐Gaussian (and potentially nonidentical) and when dependence between sensor observations can get manifested in several different nonlinear ways.