Control of Unmanned Aerial Vehicle with Wing Shape Identification using Vision System and Sensor Fusion

This thesis presents a Deflection-Detection-Vision-System (DDVS) for unmanned aerial vehicles (UAV) fixed-wing for control and navigation. This technique allows measurement of the fixed-wing shape, deflection, and identification of the aerodynamic coefficient acting on the system, using information from the stereo camera and strain gauge. The model UAV is equipped with a stereo camera fixed at the top rear end of the device and strain gauges placed in eight different points marked on the wing. Both sensors measure the deflection in chosen locations simultaneously. The DDVS performance and dynamic parameters are tested in a wind tunnel at speeds ranging from 10 km/h to 35 km/h, angles of attack (AOA), and roll angles ranging from 0 degrees to 30 degrees, respectively. An image acquisition, feature extraction, matching process, 3D reconstruction, and stereo camera calibration are presented in this thesis as a part of the proposed identification procedure. This approach measures the wing deflection at each selected point and identifies the maximum deflection location based on various aerodynamic conditions such as wind speed, AOA, and roll angle. The drag and lift forces were obtained using the wing's surface area, and the experiment shows that less force is required for lifting as the AOA increases. The DDVS was implemented in UAV and tested in the wind tunnel. Extensive experiments were conducted to determine the deflection of the wing in the function of flight parameters like angle of attack, roll angle, and flow velocity. The experimental results have shown that the integration of strain gauge and vision system sensors identify wing deflections accurately. Extensive simulation results were compared with the experimental results and demonstrated that the proposed method-based sensor fusion could be used even in the most demanding environment. The proposed control system is designed to control aircraft's attitude and velocity. A Fuzzy PID controller has been suggested, and the stability of the controller was verified numerically. The control system included a mathematical model of the aircraft that has been based on Cessna 172 aircraft. In simulation experiments, the system's stability and robustness were checked and verified.