Fully Convolutional Neural Networks for Automotive Radar Interference Mitigation

Abstract

The interest of the automotive industry has progressively focused on subjects related to driver assistance systems as well as autonomous cars. In order to achieve remarkable results, cars combine a variety of sensors to perceive their surroundings robustly. Among them, radar sensors are indispensable because of their independence of light conditions and the possibility to directly measure velocity. However, radar interference is an issue that becomes prevalent with the increasing amount of radar systems in automotive scenarios. In this paper, we address this issue for frequency modulated continuous wave (FMCW) radars with fully convolutional neural networks (FCNs), a state-of-the-art deep learning technique. We propose two FCNs that take spectrograms of the beat signals as input, and provide the corresponding clean range profiles as output. We propose two architectures for interference mitigation which outperform the classical zeroing technique. Moreover, considering the lack of databases for this task, we release as open source a large scale data set that closely replicates real world automotive scenarios for single-interference cases, allowing others to compare objectively their future work in this domain. The data set is available for download at: http://github.com/ristea/arim.

Automotive Radar Interference Mitigation with Unfolded Robust PCA based on Residual Overcomplete Auto-Encoder Blocks

“…Moreover, as we combine iterative algorithms, which provide a natural recurrent architecture, with residual overcomplete auto-encoders based on convolutional layers, we exploit the benefits of both recurrent and convolutional architectures in order to attain better results. In the experiments, we show that our approach surpasses both FCN models of Ristea et al [12]. Unlike other deep models, which only predict the amplitude, our approach is also able to estimate the phase.…”

“…A series of recent methods [9,10,11,12] rely on deep learning models to mitigate RFI. Rock et al [9] proposed a convolutional neural network (CNN) to address RFI, aiming to reduce the noise floor while preserving the signal components of detected targets.…”

“…Similarly, Fan et al [11] proposed a method based on CNNs, their contribution being that of adding residual connections, inspired by the ResNet model [14], into the architecture. In another recent work, Ristea et al [12] proposed fully convolutional networks (FCNs) trained on synthetic data samples. The FCNs, as well as other deep models, can estimate the absolute value of range profiles, but are not able to obtain the phase information.…”

A Deep-Learning Method for Radar Micro-Doppler Spectrogram Restoration

“…FCNs [ 14 ] uses FCNs to remove the interference in the FMCW signals and outputs the corresponding clean range profiles. In this approach, the main goal is to mitigate the interference, and how to restore the interfered signals with good quality is not considered.…”

“…Fan et al [ 13 ] proposed a deep residual network for narrow-band interference and wide-band interference mitigation. Ristea et al [ 14 ] proposed fully convolutional neural networks (FCNs) to remove the interference in the FMCW signals and outputs the corresponding clean range profiles. In this approach, the main goal is to mitigate the interference, but how to restore the interfered signals with good quality is not considered.…”