New procedure for disturbance-adaptive target recognition in a CFAR processor

The use of constant false alarm rate (CFAR) procedures based on ordered statistics in various model-type signal situations is studied, and it is shown that these procedures have significant advantages over previously used CFAR procedures. Three different situations are considered: the classic signal situation with a stationary disturbance in the radar picture, a transitional situation between backgrounds with very different characteristics, and a situation where two required targets are narrowly separated. Two well-known procedures for evaluating adaptive threshold values in radar problems, the cell-averaging CFAR and the cell-averaging CFAR with greatest of (GO) selection, are analyzed in detail.