Reconstruction of HRRP with missing data and noise

Abstract

This paper applies a recently developed technique for Bayesian reconstruction of Fourier pairs to reconstruct the impulse response (or the High Range Resolution Profile-HRRP) of a stepped-frequency radar target assuming missing data (backscatter), and assuming additive Gaussian noise corrupting the data. This technique is very successful at recovering HRRP of a target with significant amount of backscatter missing. The reconstructed data is used for target identification with performance nearly close to the optimal classifier.