Automatic Modulation Classification of Radar Signals Using the Pseudo Margenau-Hill Distribution

Abstract

In a non-cooperative environment, it is difficult for an electronic intelligence (ELINT) receiver to implement intra-pulse modulation classification. This paper presents an approach based on the Pseudo Margenau-Hill distribution (PMHD). The algorithm computes the PMHD of the detected radar signals and extracts the ridges. Then, it defines three characteristic features, namely, the coefficient of piecewice linear fitting, the hop number of the ridge and the coefficient of inverted V Shape and realises the automatic modulation classification (AMC) of four kinds of radar signals under negative signal to noise ratio (SNR). The probability of successful recognition (PSR) is greater than 90%.