Neural network based adaptive flight control of UAVs

The development of new control strategies for quadrotor types of unmanned aerial systems is important in meeting the existing research needs to handle undesired uncertainties for more satisfactory control systems. A Neural Network-Proportional plus Velocity (NN-PV) controller is proposed for improved stability and robustness in the presence of external disturbances and model parametric uncertainty. The experimental validation is conducted using Quanser's Autonomous Vehicles Research Studio (AVRS) for multiple-vehicle testing. The developed controller is validated through indoor flight tests. The experimental approach provides understanding and the manipulation of the controller (Angle Mode) of the QDrone. The performance of the control system demonstrates improvement of the speed and stability for the formation flight of two drones. The work presented produces a flexible, robust, and effective control system model that leads to expanded stability and reduces the effect of disturbances.