On centralized composite detection with distributed sensors
Cuichun Xu, Steven Kay · 2008
For composite hypothesis testing, the generalized likelihood ratio test (GLRT) and the Bayesian approach are two widely used methods. This paper investigates the two methods for signal detection of a known waveform and unknown amplitude with distributed sensors. It is first proved that the performance of the GLRT can be poor. Secondly, a direct way of improving the GLRT is proposed. Thirdly, an approximate Bayesian detector is derived and it is shown to be another way of improving the GLRT. Compared with the exact Bayesian approach, the proposed method always has a closed form and hence is easy to implement. Computer simulation results show that the approximate Bayesian detector outperforms the GLRT when only a few sensors receive a large signal.