Adversarial interference mitigation for automotive radar

With the massive application of radars in autonomous driving, mutual interference between radars has be-come a key issue. Over the past decades, many approaches have been proposed to solve this problem in time domain, frequency domain and space domain. These methods are mostly model-based, whose performance may suffer from inaccurate modeling. In some works, the convolutional neural networks are applied to address this problem. In this paper, we propose to mitigate automotive radar interferences using convolutional neural net-works with an adversarial framework. The performance of the model-based methods and the proposed learning-based method will be compared.