Multi-bit decentralized detection of a non-cooperative moving target through a generalized Rao test

We consider decentralized detection (DD) of an uncooperative moving target via wireless sensor networks (WSNs), measured in zero-mean unimodal noise. To address energy and bandwidth limitations, the sensors use multi-level quantizers. The encoded bits are then reported to a fusion center (FC) via binary symmetric channels. Herein, we propose a generalized Rao (G-Rao) test as a simpler alternative to the generalized likelihood ratio test (GLRT). Further, the asymptotic performance of a trajectory-clairvoyant (multi-bit) Rao test is leveraged to develop an offline and per-sensor quantizer design. Simulations show the appeal of G-Rao test with respect to the GLRT, and the gain in detection by using multiple bits for quantization.