Automatic Object Discrimination Based on Natural Resonant Features of Dielectric Coated Object

Abstract

An electromagnetic signal scattered from a radar target consists of resonance information, which can be utilized to achieve automatic target discrimination. To extract this resonant information, the short-time matrix pencil method (ST-MPM) is one of the robust approaches. The library of unique extinction pulses (E-pulses) is generated using the target’s resonant poles. Then, the quantitative discrimination measures of E-pulse discrimination number (EDN) and target discrimination number (TDN) are calculated. Recently, the radar targets are evolved with complex structures along with dielectric-coated materials, thus creating difficulty in target recognition. This communication focuses on the discrimination of dielectric-coated composite objects using extracted resonant poles and automatic discrimination measures. The demonstration of a discrimination approach is done based on simulation and measurement tests using complex scale models of aircraft. The robustness of the TDN measure is also investigated by analyzing noise-contaminated object returns for a range of signal-to-noise ratios (SNRs). This analysis shows that the TDN provides superior discrimination for all considered target sets.