Deep Learning for Classification of Mini-UAVs Using Micro-Doppler Spectrograms in Cognitive Radar

Abstract

Military radar can not only provide kinematic information (position, speed, and course) of land, sea, and air targets during day and night in all weather conditions, but also provides information about the type of target using micro Doppler properties. The micro -Doppler properties of a target are determined by the moving parts on the target body. The number, location, and type of motion of these parts are specific for a given target type and consequently the related micro -Doppler signature can be exploited for classification. Analysis of the micro -Doppler signature may provide detailed properties of rotating parts, such as the rotation rate, number of blades, and blade length. However, for operators and/or image analysts, the interpretation and understanding of radar micro -Doppler spectrograms is much more difficult and time-consuming than the analysis of optical images because of the different nature of the radar imaging principle and target scattering mechanisms. Consequently, there is a need for automatic target recognition (ATR) in radar using micro -Doppler spectrograms.