Comparison of learning models for wideband interference mitigation in automotive chirp sequence radar systems

抄録

CS (Chirp Sequence) radar plays a crucial role in the safety of autonomous driving. However, its widespread adoption increases the probability of wideband inter-radar interference, leading to miss-detection of targets. To address this problem, we utilize RNN (Recurrent Neural Network) and self-attention models to mitigate wideband interference in automotive radar systems and compare 12 different learning models in terms of SNR (Signal-to-Noise Ratio) and processing time.