Neural Network-Based Adaptive Attitude Controller for a Micro-Unmanned Aerial Vehicle

The deployment of micro-Unmanned Aerial Vehicles (UAVs) for civil applications has exponentially grown in the last decades due to their versatility, low cost and ease of use. However, their effective implementation requires an efficient attitude control system that ensures the tracking of a reference signal, even in the presence of uncertainties or external disturbances. The main aim of this paper is to present the design of a neural-network-augmented 𝐻∞ controller for a micro-UAV, to compensate for dynamic uncertainties and environmental perturbations. The effectiveness of the proposed architecture is validated through both numerical sensitivity analyses and experimental tests on a Crazyflie 2.1+ quadrotor mounted on a tilting test bench.