Micro-Doppler detection and target identification using Artificial Neural Network

Abstract

When the radar transmits electromagnetic waves, the target characteristics can be extracted from the received echo signal. During the reception process, the radar carrier frequency will be shifted due to the Doppler effect of moving targets. If a specific target makes any vibration or rotation, an induced frequency modulation on the received echo signal will be occurred generating side-bands around the Doppler frequency; this is called the micro-Doppler (m-D) phenomenon. To analyze and separate the m-D signature from the received signal, some extracted features techniques such as Fast Fourier Transform (FFT), Time-Frequency Representation (TFR), and Wavelet Transform (WT) can be used. In this paper, the identification of the m-D has been achieved using a supervised Artificial Neural Network (ANN) identifier. The input of ANN identifier are a group of extracted features related to the received signal. The performance of the ANN identifier were tested, and the accuracy obtained has been ranging between (82.5%) and (100%).