Lower Bounds on Pose Estimation with High Range - Resolution Radar

High range-resolution radar (HRR) is a useful tool for pose estimation: the problem of determining the orientation of a target object. Using a blocks-world tank as target, HRR models for return data for both rough and specular surface reflections are determined. The profiles are used to find Cram6r-Rao type bounds on mean-squared error performance of any unbiased estimator which uses HRR data. The goal is to determine the effect on these bounds of the key parameters of the system: signal-to-noise ratio (SNR), target size, range bin width or range-resolution, and target reflectivity. Results are compared with minimum mean-squared error estimates obtained via simulation. Thesis Supervisor: Jeffrey H. Shapiro