Classification of UAV-to-ground vehicles based on micro-Doppler effect and bispectrum analysis

Abstract

Vehicles such as armored cars and tanks have a big threat due to their flexibility and lethality in modern wars. In order to destroy them without casualties, the unmanned aerial vehicles (UAVs) are widely used in local high-precision strike. For the purpose of best attack plan, it is necessary and significant to find a way that can distinguish ground wheeled vehicles and ground tracked vehicles from the UAV with high accuracy. In this paper, a classification method based on micro-Doppler effect and bispectrum analysis is proposed. Firstly, models describing relationship between ground vehicles and the UAV are established to derive radar echo signals. Secondly, bispectrum is utilized to analyze echo signals and diagonal slice of the bispectrum is obtained by calculating the third-order accumulation of echo signal. According to the difference of ground vehicles, three features are extracted. Thirdly, these features are sent to support vector machine for classification. Results using simulated data and measured data in different cases prove the effectiveness and robustness of proposed method. Comparison with current methods also verifies the superiority of method in this paper.