Optimization of a serial distributed detection system with 2 bits communication constraint

In this paper, we consider a serial distributed detection system of two sensors in which the first sensor is allowed to communicate 2 bits to the second sensor (global detector). We use the J divergence to optimize the system and the optimization is obtained by varying the binary thresholds of the first sensor to obtain the maximum value of the global probability of detection for a specified global probability of false alarm. The numerical results obtained for the J divergence does not satisfy the expected performance improvement. Motivated by this, we propose a new more direct optimization procedure. The numerical results indicate the performance superiority of the proposed procedure over that of the J divergence method. Moreover, the performance of J divergence and the proposed method is better than that of the distributed system in which 1 bit hard decision is communicated between sensors.