Airflow-based odometry for MAVs using thermal anemometers

This article concerns airflow-based odometry for estimating MAV flight speed from airflow measurements provided by a set of thermal anemometers. Our approach relies on Gated Recurrent Unit (GRU) based deep learning to extract features noisy and turbulent measurement signals triaxial anemometers, in order establish the underlying mapping between speed. The proposed solution is validated multi-rotor MAV. results show that GRU-based model can effectively noise perform denoising, compensate induced velocity effects along propellers’ rotation axis. As consequence, robust prediction performed, including during takeoff landing induce ground strong variations vertical airflow.