Classification of UAV and Ground Targets By Micro-Doppler Signatures Based on PD Surveillance Radar

As typical of low flying altitude, small size and slow flying speed (LSS) targets, the detection and recognition of mini-sized unmanned aerial vehicle (U A V) by surveillance radar is of great significance. However, it is difficult to realize classification of UA V and ground targets based on pulse Doppler (PD) surveillance radar because of low resolution and short dwell time. Based on the method of multiple micro-Doppler signatures (m-Ds) association, classification of UA V and non-UAV targets is considered and studied in this paper. We first focus on the micro-Doppler signature of UAV targets by PD radar, including the echo model and the micro-Doppler spectrum distribution. To verify the proposed method, we investigate UAVs, human and vehicle using the developed PD surveillance radar in a series of experiments. And seven key features are extracted from both frequency domain and time-frequency domain. The experiment results show that the classification accuracy is found to be more than 90%.