Development of An Autonomous Precision Landing System for Quadrotor Unmanned Aerial Vehicles (UAV)

Unmanned aerial vehicles (UAVs) are used in various fields, and the most important aspect of UAVs' automatic operation is landing. Currently, most UAVs use only data or signals received from satellites for automatic landing. However, using satellite data can be quite inaccurate, especially if the UAV is landing in an area where satellite signals are unstable. Therefore, most users of UAVs prefer to use the infrared light landing method, which involves using a camera with a lens that can detect infrared light on the UAV and a device that emits infrared light placed at the landing pad. This method provides less error than using satellite data, but the equipment is relatively expensive such as the cost of an infrared light landing pad is up to 100 times higher than that of a computer vision landing pad, and the cameras used are also more expensive. In this research, we chose to use computer vision for automatic landing. This involves placing a camera on the UAV and placing 2 AprilTags on the landing pad by nesting a small AprilTag at center of a big AprilTag. This allows the UAV to land accurately. Through testing, this method resulted in an average error of only 0.1158 meters.