Deep Learning and Information Geometry for Drone Micro-Doppler Radar Classification

In this work, we build dedicated learning models for micro-Doppler radar time series classification. We develop both deep temporal architectures based on time-frequency representations, and also directly study the signal's underlying statistical Gaussian process using Information Geometry on Riemannian manifolds by developing and improving symmetric positive definite (SPD) neural networks. We also propose the aggregation of all proposed models in a single, highly performing classification pipeline.