Ultra-wideband radar noise reduction for target classification

Abstract

The authors describe a calibration procedure for ultra-wideband (UWB) radar systems in order to measure the frequency-dependent transfer function or radar cross-section, respectively, of objects with a high computational efficiency. The benefit of using UWB pulses is that only one measurement is required for covering several GHz. However, the signal processing steps contain a deconvolution operation producing spike artefacts. They were successfully removed by the calculation of frequency-dependent signal-to-noise ratio and a following Wiener filtering. The proposed method was verified by some elementary targets of which the size of the objects were determined.