Autonomous landing system design of rotary wing unmanned aerial vehicles on a moving platform

Unmanned Aerial Vehicles (UAVs) have created a significant transformation in technology and engineering, finding extensive use in various civil, military, and industrial applications. They are particularly effective in tasks such as reconnaissance, observation, cargo transportation, and search and rescue operations. However, to enhance the effective use of UAVs and enable them to perform more challenging tasks, the development of autonomous landing and takeoff capabilities is crucial. This study presents a system developed for UAVs to autonomously land on a moving platform, supported by position information, Kalman filtering, and image-based object detection methods. In this study, the system developed for UAVs to autonomously land on a moving platform in a simulation environment utilizes not only the UAV's own position data but also the position, speed, and orientation data of the moving platform. The position information of the landing pad was obtained using a specially trained YOLO V8, enabling high-precision landing. Four innovative scenarios were created to measure and improve the system's performance. These scenarios demonstrate that the UAV can autonomously land on a moving platform after various measurements in different altitudes and regions. The system's performance under side winds with variable speed and direction, as well as under limited visibility conditions due to fog, has been tested through simulations. Additionally, in a scenario with multiple landing pads, the UAV successfully selected the nearest ground vehicle as its landing target at the beginning of the operation. The effective area of the landing system is expanded by the GPS location provided by the moving platform. A Kalman filter, integrated with the GPS position of the platform and the orientation angle taken from the Inertial Measurement Unit on the UAV, has been used. A meeting point in the direction of movement of the platform is determined with Kalman filter, and thus a more efficient trajectory is obtained. Compared to similar studies in the literature, the novelty of this work lies in developing a new approach for approach and landing trajectories that involve altitude changes and achieve the target via a shorter route. The developed system offers integration possibilities with various UAV and ground vehicle platforms, providing solutions suitable for different mission requirements.