Vision-based multi-sensor fusion for robust unmanned aerial vehicles autonomous navigation

The primary objective of this dissertation is to propose a robust vision-based multi-sensor fusion navigation system for lightweight unmanned aerial vehicles to navigate and perform complex tasks in a challenging environment. In recent years, the demand for autonomous drones has grown exponentially. Such applications can be the day to day routine goods delivery or inspection work in an area that humans can't reach. The proposed navigation system can run in real-time, low cost and with a high level of robustness. The proposed methods use the camera as the primary sensor to both odometry and mapping. Then the multi-sensor fusion is done to enhance the odometry reliability of the UAV system.Doctor of Philosoph