A sensor network is considered where a sequence of random variables isobserved at each sensor. At each time step, a processed version of theobservations is transmitted from the sensors to a common node called the fusioncenter. At some unknown point in time the distribution of the observations atall the sensor nodes changes. The objective is to detect this change indistribution as quickly as possible, subject to constraints on the false alarmrate and the cost of observations taken at each sensor. Minimax problemformulations are proposed for the above problem. A data-efficient algorithm isproposed in which an adaptive sampling strategy is used at each sensor tocontrol the cost of observations used before change. To conserve the cost ofcommunication an occasional binary digit is transmitted from each sensor to thefusion center. It is shown that the proposed algorithm is globallyasymptotically optimal for the proposed formulations, as the false alarm rategoes to zero.
展开▼