Modal ordered shape factors as radar feature set

Abstract

Abstract The characteristic polarisation states form a second layer feature set by reflecting shape attributes about that target, enabling better identification performance of the resonance signature. These shape factors reflect a structure's curvature extent, dihedral degree between corners, and the axial ratio between principal axes by determining two characteristic angles associated with the null polarisation state and a ratio of the optimum maximum and minimum received powers, respectively. However, the accuracy of the shape factors degrades with a poorly estimated resonance signature caused by noise, missing resonance due to occlusion or ambiguity in late time onset. Thus, the authors aim to reduce the effect of these problems using an ensemble average to filter noise and enhance the signal strength, properly selecting a modal order to ensure modal consistency of the signature and decay sum (DS) to select the late time onset properly to avoid missing resonance within the polarisation matrix. Finally, a paradigm of two jetfighters validated the factors' discriminative potential across an azimuth plane of low depression angle. The results showed that a DS around 0.4 improves the estimated factors over most resonance modes and azimuth directions. At most target aspects, the first‐order shape factors consistently predicted a dominant parallel wedge‐shaped structure, while the second‐order shape factors consistently predicted a trough‐shaped structure; finally, the third‐order factors revealed wedge‐shape attributes at forward look aspects but trough‐shaped attributes at backward look aspects.

Keywords

• polarimetric radar
• radar target recognition