A novel variational autoencoder based radar signal reconstruction algorithm using polluted data

In the transmission process, radar signal could be polluted easily, and the radar individual recognition process will be undermined by the polluted data. Some existing algorithms so far use the image denoising method to deal with this issue while ignoring the essential characteristics of the signal, and thus the noise reduction effect is poor. To address this issue, a novel variational autoencoder based radar signal reconstruction algorithm is proposed in this paper to reconstruct high quality data from the polluted ones by compressing the one-dimensional polluted signal and learning the essential characteristics. The experiments prove that the algorithm indeed improves the quality of the reconstructed data compared with the polluted signal.