Compressed Sensing for UWB Radar Target Signature Reconstruction

Ultra-wideband radar target impulse response is characterized by few significant scatterers that can be directly associated with the target geometry such as wing tip, engine inlet, etc. Although there are few scatterers of interest, the UWB nature of the radar sensor requires extensive frequency domain measurements or highly sampled time domain signatures. This requirement is simply a consequence of Nyquist sampling criterion despite the limited number of scatterers along the range (fuselage) of the target. This paper employs the relatively new concept of "compressed sensing" for generating an UWB target signature with as little measurements as possible. The paper applies the concept of compressed sensing to synthetic target models then extends the same approach to real radar signatures recorded in a compact range environment. The results indicate the feasibility of reducing the number of samples needed to accurately characterize and identify target scattering centers.