Decentralized sequential detection with sensors performing sequential tests (Q1894654)

In decentralized sequential detection each one of a set of sensors receives a sequence of observations about certain hypotheses \(H\); in the present case, the authors consider two sensors and two different hypotheses. Furthermore, it is assumed that each sensor first performs a sequential test on its observations and arrives at a final local decision \(u_i\), \(i=1,2\); subsequently, the local decisions are sent to a fusion center for a common goal represented by a common cost function \(W(u_1, u_2; H)\). It is shown that the optimal sensor decision functions can be found in the class of generalized sequential probability ratio tests (GSPRT) with monotonically convergent thresholds. A procedure for the numerical computation of the optimal thresholds is given. Numerical results are presented when the sensor observations are normally distributed under each hypothesis. Moreover, the performance of the optimal GSPRTs is compared with the best performance that is obtained when the sensors are restricted to using SPRTs.