Classification of Airborne Radar Signals based on Time-Frequency Features using Wigner-Ville Distribution

Abstract

This paper presents a classification system for airborne radar signals using Wigner-Ville distribution (WVD) and rule-based classifier for use in the field of electronic warfare (EW) for electronic intelligence gathering. The signals considered in this paper are mostly of multi-group low probability of intercept (LPI) capabilities of phase and frequency modulation origin. The WVD used in this paper was altered using two window functions in the time-lag domain in order to counteract the shortcomings of the normal WVD. The classifier was based on time, frequency and phase analyses carried out in order to estimate important features for the classifier rules. Performance analysis was carried out in order to determine classification accuracy. Results obtained showed a classification accuracy of 100% at signal-to-noise ratio (SNR) equal to or greater than 1 dB. Computational complexity analysis of the methodology used showed a highest order of three, similar to previous related paper.

Keywords

• electronic warfare (ew)
• low probability of intercept (lpi)
• rule-based classifier
• signal-to-noise ratio (snr)
• wigner-ville distribution (wvd)