Radar Antenna Scan Pattern Intelligent Recognition Using Visibility Graph

Radar antenna scan pattern (RASP) reconnaissance is a major problem in electronic warfare (EW). The RASP exerts a considerable influence on target identification, jamming decision making, and electronic support measures and thus plays a critical role in modern electronic warfare. A visibility graph (VG) is a tool for converting a time series into complex graphs with excellent noise immunity. This paper proposes a novel method for the intelligent recognition of the RASP based on the VG, including the circular, sector, helical, raster, conical, phased array, phased array azimuth and circular elevation scans. The changes in the signal amplitude received from the EW receiver are determined. Moreover, the related features are extracted from the VG and utilized to classify the RASPs. The comparison experiments performed with different classifiers, such as machine learning, neural network, and deep learning, confirm that the proposed method can improve the robustness of the recognition rate to the noise and recognition accuracy.