State-space approach to decentralised detection

A discrete-time decentralised detection problem is considered where the data generating processes are represented by state-variable random process models. The optimum detector is constructed by the computation of the local likelihood ratios at each sensor in real time, and the subsequent transmission of these ratios to the fusion centre. Decentralised detection based on the fusion of the likelihood ratios is expected to perform as well as centralised detection, because there is no information loss due to intermediate decision-making.