Statistical analysis of hypersonic glide vehicle radar cross section

Abstract

AbstractThe capability to design, manufacture and test Hypersonic Glide Vehicles (HGVs) has been demonstrated by a number of nations, and they are increasingly forming part of military inventories, potentially offering capabilities highly unique to this technology. This article reports the simulated Monostatic Radar Cross Section of a generic HGV in five frequency ranges, HF, VHF, UHF, L, and S‐bands associated with different radar types. Full spherical datasets of complex co‐ and cross‐polar data are synthesised so that backscatter resulting from illumination by r.f./microwave energy of linear or circular polarisation can subsequently be computed from the raw dataset. Circular polarisation is commonly employed by ground‐based Ballistic Missile Early Warning Systems and Space Object Surveillance and Identification radars to avoid polarisation mis‐match losses resulting from ionospheric Faraday rotation effects. The data was generated using Ansys' Finite Element Solver at 10, 150 and 430 MHz, with the Geometric Optics/Physical Optics based SBR+ solver employed for 1.3 and 3 GHz data. All data was produced at below the Nyquist sampling interval relevant to the target's electrical size. These datasets were then imported into a Matlab routine which extracted data over limited angular ranges associated with the likely radar line‐of‐sight in particular scenarios, typically having a standard deviation of ±10° about the direction of flight, applying either a Gaussian or Uniform sampling distribution as part of a Monte Carlo analysis. These extracted data were then used to form histograms giving the probability of sampling particular RCS values. Probability density functions and cumulative distribution functions were then fitted, to aid in the representation of statistical target fluctuations for each band and angular sampling range. The HGV exists in either the ‘Rayleigh’, ‘resonance’ or ‘optical’ scattering regimes, depending on its relative electrical size. The results suggest that for this target shape at HF and VHF cases a simple Swerling 0 (fluctuation invariant) approximation is adequate in most instances, whilst a Gamma distribution may be applied for UHF band cases. At L and S‐band a Beta distribution was found to provide a good fit to the available data.