ADAPTIVE DETECTION OF STATIONARY GAUSSIAN SIGNALS AGAINST A NORMAL NOISE BACKGROUND, WITH A CONSTANT FALSE-ALARM RATE

Abstract

Purpose: Efficiency analysis of the Cell-Averaging Constant False Alarm Rate processor (CA CFAR-processor) as applied to detection of stationary Gaussian signals against a normal noise background with unknown and/or varying from scan to scan power. Design/methodology/approach: Standard methods of the theory of optimal filtration and statistical signal processing are used to calculate the true detection probability and false-alarm rate. Findings: Analytical expressions have been derived for the scaling factor which ensures a constant level of the false-alarm rate, as well as for the true detection probability in dependence on the number of the reference cells and signal-to-noise ratio. It is shown that for efficient application of the given algorithm, the number of the reference cells should be 20 to 30, depending on the signalto-noise ratio μ. In this case, the amount of loss in the signal-tonoise ratio does not exceed 1 to 2 dB as compared with the situation where the noise power is a priori known and invariable. With μ≥30 dB the amount of loss proves to be negligibly small and the need in adaptation vanishes. Conclusions: The results obtained testify to the efficiency of application of the CA CFAR processors to detection of targets corresponding to Swerling model 1 against a normal noise background with unknown power associated with clutter and/or scattering from irregularities of the propagation medium.

Keywords

• target detection
• false alarm rate
• detection threshold
• signal-to-noise ratio
• cell averaging