Paper Info
Deep radar detector
D Brodeski, I Bilik, R Giryes
2019 IEEE Radar Conference (RadarConf), 1-6, 2019.
Reviews
From
yassendobrev, rated
4/5.
Oct. 28, 2021.
The paper describes an approach replacing (and outperforming) a 2D CFAR by a convolutional neural net and the Bartlett Beamformer by a few fully-connected layers. The network was trained using augmented chamber data. It's unclear to me if the range / Doppler / azimuth / elevation output is limited to a grid and if the network supports more than one target per grid cell.
Deep radar detector
D Brodeski, I Bilik, R Giryes
2019 IEEE Radar Conference (RadarConf), 1-6, 2019.
Reviews
From
yassendobrev, rated
4/5.
Oct. 28, 2021.
The paper describes an approach replacing (and outperforming) a 2D CFAR by a convolutional neural net and the Bartlett Beamformer by a few fully-connected layers. The network was trained using augmented chamber data. It's unclear to me if the range / Doppler / azimuth / elevation output is limited to a grid and if the network supports more than one target per grid cell.