A Convolutional Neural Network-Based Method for Discriminating Shadowed Targets in Frequency-Modulated Continuous-Wave Radar Systems

Sensors

Maurizio Valle¹  Christian Gianoglio¹  Ammar Mohanna¹  Ali Rizik¹ 

[1] Department of Electrical, Electronic and Telecommunication Engineering and Naval Architecture (DITEN), University of Genoa, Via Opera Pia 11, 16145 Genoa, Italy;

关键词: radar;    shadow effect;    machine learning;    CNN;    transfer learning;   

DOI  :  10.3390/s22031048

来源: DOAJ

【 摘 要 】

The radar shadow effect prevents reliable target discrimination when a target lies in the shadow region of another target. In this paper, we address this issue in the case of Frequency-Modulated Continuous-Wave (FMCW) radars, which are low-cost and small-sized devices with an increasing number of applications. We propose a novel method based on Convolutional Neural Networks that take as input the spectrograms obtained after a Short-Time Fourier Transform (STFT) analysis of the radar-received signal. The method discerns whether a target is or is not in the shadow region of another target. The proposed method achieves test accuracy of 92% with a standard deviation of 2.86%.

【 授权许可】

Unknown