An improved reduced-rank CFAR space-time adaptive radar detection algorithm

摘要

The CFAR test developed in previous work is a normalized ratio test for signals in nonwhite Gaussian noise. However, in the airborne radar environment, the noise consists of strong interference and a relatively weak thermal noise, in the case of a large interference-to-thermal noise ratio, this test can be simplified to the reduced-rank CFAR test developed previously, which operates in an interference-free subspace without the need for matrix inversion operations. This test is extended in this paper to one that includes both the primary and secondary data as defined by Bose and Steinhardt (see ibid., vol.43, p.2164-75, 1995), it is also shown that this test can be modified to obtain a dramatically improved performance. A much smaller amount of sample data is needed in this new improved algorithm to achieve a given probability of detection than is required by this test. Finally, the performance of this new reduced-rank CFAR test statistic is analyzed, and a simulation is performed for an example scenario in order to validate the theoretical results.