Ultrawideband inverse scattering method for resonance region target recognition: Application to small-scale airplane targets with measured data

This paper develops an inverse scattering method to determine the characteristics of resonance region targets by using ultrawideband scattering signals. The method mainly uses multiple time-domain signals for each target, and they are processed to give feature vectors in frequency domain. These vectors contain target-specific information related with characteristics (dimensions of substructures, material properties, etc.) of the known targets. The target recognition is employed by using test (unknown) target´s scattering signal. This time-domain signal is processed in an inverse way to get a frequency-domain test vector. Then, the characteristics of test target are assigned as these of one of the known targets with a basic template-matching approach between test vector and feature vectors of these targets. The method is demonstrated for three realistic small-scale airplane. For an ultrawideband system of 0.7-18 GHz, more than 85 percent accuracy has been obtained even under moderate signal-to-noise (SNR) ratio of about 10 dB.