A New Radar Detector in Unknown Signal and Clutter Environment

In this paper, we introduce a new detector in the absence of any statistical knowledge of the fluctuating signal; perhaps a weak signal, and clutter. We use a few sampled fractional moments (FM) to construct the maximum entropy (MAXENT) probability density function (PDF) estimation. These moments; i.e., their fractional orders, are obtained from the observed sample variates. Using the fractional moments instead of the integer moments the estimated PDF is quite close to the true PDF. The test statistics is a fractional polynomial of very low order of the received samples. We consider the following target fluctuating models, swerling, lognormal, and Rician. We also consider the clutter to follow low and heavy tail models; i.e. Rayleigh, and lognormal