eduzhai > Applied Sciences > Engineering >

Estimating Magnitude and Phase of Automotive Radar Signals under Multiple Interference Sources with Fully Convolutional Networks

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 12 pages

Abstract: Radar sensors are gradually becoming a wide-spread equipment for roadvehicles, playing a crucial role in autonomous driving and road safety. Thebroad adoption of radar sensors increases the chance of interference amongsensors from different vehicles, generating corrupted range profiles andrange-Doppler maps. In order to extract distance and velocity of multipletargets from range-Doppler maps, the interference affecting each range profileneeds to be mitigated. In this paper, we propose a fully convolutional neuralnetwork for automotive radar interference mitigation. In order to train ournetwork in a real-world scenario, we introduce a new data set of realisticautomotive radar signals with multiple targets and multiple interferers. To ourknowledge, this is the first work to mitigate interference from multiplesources. Furthermore, we introduce a new training regime that eliminates noisyweights, showing superior results compared to the widely-used dropout. Whilesome previous works successfully estimated the magnitude of automotive radarsignals, we are the first to propose a deep learning model that can accuratelyestimate the phase. For instance, our novel approach reduces the phaseestimation error with respect to the commonly-adopted zeroing technique byhalf, from 12.55 degrees to 6.58 degrees. Considering the lack of databases forautomotive radar interference mitigation, we release as open source ourlarge-scale data set that closely replicates the real-world automotive scenariofor multiple interference cases, allowing others to objectively compare theirfuture work in this domain. Our data set is available for download at:this http URL.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×