Classification of LPI radar signals using spectral correlation and support vector machines

Abstract

In modern radar systems, low probability of intercept (LPI) waveforms are used to make detection by a potential adversary difficult. This is accomplished using wideband waveforms, frequency hopping, and continuous waveforms (FMCW) to reduce the signal profile. The low signal profile of the LPI signal enables the radar to perform detection and or target tracking while the target remains unaware. Several modulation techniques such as polytime codes, polyphase codes, FSK, and FMCW are used to produce LPI signals for transmission. This paper looks at the ability of spectral correlation along with a support vector machine in order to automatically classify the different LPI signal types in a non-cooperative environment.