Consistency research between statistical processing and array processing for detection of narrowband signals in presence of a coherent interferer

The detection problem of a narrowband signal in the presence of a single coherent interferer is studied in the perspectives of statistical and array signal processing respectively, corresponding with the method of Generalized Likelihood Ratio Test (GLRT) and the way to change the beamformer to derive for the suboptimum detector. The consistency between statistical processing and array processing for the given problem is proved via theoretical derivations and numerical simulations in this paper. The noncentrality parameter of the statistical distribution, which the test statistic derived in the statistical processing follows, is positively correlated to the array gain. The detectors for the given problem, derived respectively in two different processing views, thus converge to be one. Its detection performance is studied by computer simulations, considering the target of varied bearings and different element number of uniform linear array (ULA) respectively, both with a constant interference bearing. From the simulation results and their analysis, the conclusions can be drawn: (1) the detection performance will increasingly decrease as the target approaches the interferer increasingly when the target signal incident between the upper critical bearing and the lower one; otherwise, the detection performance will remain relatively constant; the formula for the two critical bearings is derived specifically for the ULA; (2) the bearings of the target and interference can be relatively closer without worsening the detection performance when more elements used.