Ordered statistic CFAR technique - an overview

Adaptive radar target detection in a noise or clutter environment is a very important device in each radar receiver. In almost all detection procedures the received echo signal amplitude is simply compared with a certain threshold. The main objective in target detection is to maximize the target detection probability under the constraints of very low and Constant False Alarm Rate (CFAR). The noise and clutter background will be described by a statistical model with e.g. independent and identically Rayleigh or exponentially distributed random variables of known average noise power. But in practical applications this average noise or clutter power is absolutely unknown and can additionally vary over range, time and azimuth angle. Therefore this paper describes some of the so-called range CFAR techniques for several different background signal situations in which the average noise power and some other additional statistical parameter are assumed to be unknown. All range CFAR techniques combine therefore an estimation procedure (to get precise or estimated knowledge about the noise power) and a decision step by applying an amplitude threshold to the echo signal amplitude inside the test cell. This general detection scheme has been analysed in many research activities and a large effort has been spent on this topic. The objective of this paper is to give a short characterisation and technical comparison of some of these important range CFAR detection schemes.