无线电工程2024，Vol.54Issue(1)：63-70,8.DOI:10.3969/j.issn.1003-3106.2024.01.009
基于注意力机制和复数卷积循环网络的汽车雷达干扰抑制
Automotive Radar Interference Suppression Based on Complex Convolution Recurrent Network with Attention Mechanism
摘要
Abstract
With the development of autonomous driving technology,more and more vehicles are equipped with on-board radars.But the on-board radars of different vehicles may interfere with each other,leading to the appearance of false targets or the increase of floor noise,which reduces the detection performance.A Deep Complex Convolution Recurrent Network with Attention(DCCRN-Attention)is proposed to solve the problem of mutual interference between automotive radars,and interference suppression is realized in frequency domain.The proposed model uses complex network to combine the real and imaginary parts of the signal for feature learning,which can simultaneously predict the amplitude and the phase of the target after interference suppression.And the model introduces the attention mechanism in the skip connection to focus on the more important feature information and suppress the irrelevant information.The experimental results show that the proposed model can effectively suppress the interference,improve the Signal to Noise Ratio(SNR)of the targets,and outperform the baseline methods in the evaluation indexes.关键词
汽车雷达/干扰抑制/深度复数卷积循环网络/注意力机制Key words
automotive radar/interference suppression/DCCRN-Attention/attention mechanism分类
信息技术与安全科学引用本文复制引用
吴秋雨,高勇..基于注意力机制和复数卷积循环网络的汽车雷达干扰抑制[J].无线电工程,2024,54(1):63-70,8.基金项目
四川大学科研项目资助(0020505516013)Funded by Science and Research Program of Sichuan University(0020505516013) （0020505516013）
