Binary Stokes vector representation of aircraft in the low‐resolution radar context

Abstract

This study proposes a low-band scheme utilising the polarisation dominance of the target's resonance regions in the frequency domain to recognise a fighter-sized aircraft. The methodology implements a binary decision tree based on three polarisation power terms, namely Stokes variables, to determine the polarisation dominance of six directions, and subsequently, predict the canonical shape of the target substructures. The decision tree comprises three inequalities of the Stokes variables subject to an orthogonality threshold to reflect the dominant polarisation directions as a function of the target resonance region. The simulation data for two fighter-sized aircraft illustrate the feasibility of the method by comparing the differences (or similarity) in the target decision outcomes. The results indicate that resonance-scattering region lies within the frequency band of 9–200 MHz where thin-wedge scattering is prominent within the frequency band of 9–30 MHz, whereas the corner (dihedral) scattering becomes dominant within the band of 110–170 MHz, mainly when the radar illuminates the target sides aspect. The method was robust to the noise and target aspect change using the paradigms of the MIG29 and F16 aircraft models.