Applications of nonlinear filtering in radar CFAR detection

The primary concern in radar constant false alarm rate (CFAR) design is the estimation of background noise distributions of known parametric form, but with unknown parameters. Most systems employ an adaptive detection threshold, using an estimate of the noise level to maintain a CFAR. Various nonlinear filtering algorithms, notably, order statistic (OS) filters, are used as the CFAR estimator. The authors review the application of such nonlinear filtering techniques, progressing from Gaussian to non-Gaussian situations. They present a new method for non-Gaussian clutter.<
>