Experimental investigation of system identification techniques for UAVs

• Main Author: Zhang, Huaiyu
• Other Authors: Basman Elhadidi
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Aeronautical engineering::Aircraft; Engineering::Mechanical engineering::Mechanics and dynamics
• Online Access: https://hdl.handle.net/10356/157655
• Institution: Nanyang Technological University

UAVs are known for their remarkable performance and flexibility compared to their manned counterparts. As most UAVs are highly automated and rely predominantly on their flight control systems to maintain stability during flight, it is paramount to identify their aerodynamic derivatives, which are associated with stability and control responses of the aircraft. Doing so will facilitate structural design of the UAVs, as well as the development of their flight control systems. This final year project report proposes an innovative experimental methodology for the evaluation of aerodynamic derivatives of UAV models. The project takes both theoretical and experimental approaches to obtain the model’s pitching moment and normal force derivatives. The theoretical approach looked into flight dynamics and control theory, from which theoretical value of aerodynamic derivatives could be calculated. On the other hand, the experimental approach involved design and fabrication of a UAV test model, which was integrated with IMU and load cell. The experiment was conducted in two phases in the form of wind tunnel testing, with the first phase designed as free pitching motion of the model with its pitch rate recorded. This allowed the evaluation of pitching moment derivatives through parameter identification with the aid of MATLAB Curve Fitting Toolbox. Natural frequency and damping ratio could also be obtained through this process. The second phase of the experiment measured the normal force experienced by the model with a load cell, so that normal force derivatives could be determined by performing system identification with MATLAB SI Toolbox. The experiment results were compared against theoretical values to evaluate the accuracy of the experimental methodology proposed. The values obtained through parameter and system identification were of close resemblance to theoretical calculations, with minor differences in the transfer function model of normal force derivatives. The scope of future works could potentially be extended to study UAV models with more complicated structures and multiple degrees of freedom.