Subspace Selection for Quadratic Detector of Random Signals in Unknown Correlated Clutter
IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
| Hi-index | 35.68 |
All of the conventional CFAR detection algorithms that use space-time processing involve a time-consuming matrix-inversion operation. Based on today's technology, this computational complexity sometimes makes the full-rank solution difficult to realize. In this correspondence, a CFAR detection algorithm, which does not need a matrix inversion, is developed by an adaptation and extension of Hotelling's principal-component method studied recently by Kirsteins and Tufts (1994). Finally, the performance of the new CFAR test statistic is analyzed, and the effect of the rank reduction on performance is evaluated for an example scenario