U-Net-Based Automotive Radar Target Detection and Recognition
Jamal Kazazi, Seyyed Mohammad Matin AleMohammad, Mahmoud Kamarei · 2024
In this paper, we have designed a CNN-based neural network, to detect targets using a FMCW radar more accurately than traditional methods. We have also shown that using this approach will enable us to not only detect targets but also it can categorize them, by their size, with high accuracy. Training of such a network requires a large amount of data, so one of the most challenging parts of this research was to simulate and generate suitable radar data to train the network. The detection rate of targets is reported for different SNRs and it increased by 40 percent on average, in comparison with CFAR methods.