Multiple-sensor distributed detection systems with data fusion

This short paper presents a more general formulation to evaluate the global performance improvement in decentralized multiple-sensor distributed detection systems with data fusion. Furthermore, a comparative study of the global performance improvement in centralized and decentralized approaches is discussed in case of Gaussian distributed observations. The performance loss due to the decentralized approach is also discussed. The needed number of sensors, in the decentralized approach to achieve a performance non inferior to the centralized approach, is also obtained. We consider more complicated examples by considering a large number of sensors to characterize the performance trade-off between the centralized and decentralized approaches. Then the communications requirements can be compared to choose a preferred architecture.