Vision-based Autonomous Guidance Approach for a Nano Unmanned Aerial Vehicle Rotorcraft Towards Indoor Flight Environment

Autonomous flight remains a major challenge for Nano Unmanned Air Vehicles (NUAVs). This research presented a light-weight vision-based navigation system which is suitable for NUAV rotorcrafts with restrictively limited payload capacity and computing resources. A visual-inertia compensation approach was proposed to obtain accurate indoor navigation and guidance for the homemade NUAV which weighting less than 50 g. The proposed visual guidance algorithm have been successfully recognized targets such as windows, gates, and narrow-corridor with 25 Hz guidance command output. Flight experiments shown that the NUAV rotorcraft can follow the path planning to achieve autonomous flight, and guide through gates and windows within GPS-denied indoor environment.