Target Identification Using Global and Local Features.
Abstract
This report is concerned with input-output from an electromagnetic process. The input is a component of the field generated by the experimenter. The output is a component of the field scattered by an object. The goal is to determine, from the data, characteristics of the scatterer. We consider both complex exponential (global) and ray-optic (load) modeling. We report results from the both simulated data and experimental data obtained on the transient ranges at NOSC and Lawrence Livermore National Laboratory. In complex exponential modeling, we conclude that pole identification can be accomplished to a high degree of accuracy. Relating these data poles to complex resonances of the scatterer, however, is limited to high-Q cases in a low noise environment. Our results in ray-optic modeling are preliminary. We find serious difficulties in discriminating between smooth objects with the same radius of curvature and between identifical objects with different curvatures. Originator-supplied keywords: Target identification, Target classification, System identification, Parametric modeling, Pole-zero modeling, Complex resonance identification.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 11, 1985
- Accession Number
- ADA153967
Entities
People
- D. G. Dudley
Organizations
- University of Arizona