Parametric representation of radar signatures using three-dimensional augmented state space approach

In this paper, a three-dimensional (3-D) augmented state space approach (ASSA) is proposed for parametric representation of the radar signatures collected on targets. The 3-D radar signatures obtained from different azimuth and elevation angles are regarded as the summation of returned signatures from discrete scattering centers which are characterized via pairing the amplitudes and poles in three orthogonal plans (down range, azimuth and elevation dimension). A single augmented Hankel matrix is developed for pole estimation of the down range dimension. Multiple range search scheme, which translates the large matrix composed of all 3-D radar signatures into multiple smaller pairing matrices, is used for parameter estimation of the other two dimensions. Effectiveness of the proposed approach is tested on two numerical data, i.e. signatures composed of scattering points as well as the signatures of a sphere tipped cone-cylinder-frustum combination model. The result demonstrates that 3-D ASSA could provide robust and accurate parametric representation results in the presence of noise. In addition, it is also suitable for signal reconstruction with high compression ratio.