Target characterization is an important step in many defense missions, often relying on fitting a known target model to observed data. Optimization of model parameters can be computationally expensive depending on the model complexity, thus having models that both describe the data well and that can be efficiently optimized is critical. This work introduces a class of radar models that can be used to represent the radar scattering response of a target at high frequencies while also enabling the use of gradient-based optimization.
@article{arxiv.2206.02075,
title = {Differentiable Point Scattering Models for Efficient Radar Target Characterization},
author = {Zachary Chance and Adam Kern and Arianna Burch and Justin Goodwin},
journal= {arXiv preprint arXiv:2206.02075},
year = {2022}
}