Dynamic Modelling and System Identification of a Quadrotor Unmanned Aerial Vehicle Using Gyroscopic Based Test Stand

Abstract

System identification plays a crucial role in developing accurate mathematical models for dynamic systems, which are essential for designing high-performance controllers. This study presents a novel gyroscope-based dynamic test rig developed specifically for quadrotor-type unmanned aerial vehicle (UAV) system identification. Unlike conventional outdoor testing, the proposed platform provides a controlled, repeatable, and vibration-minimized environment for capturing roll, pitch, and yaw dynamics, ensuring high data fidelity and safety. Using Mission Planner and LabVIEW grey box optimization, the research successfully identified continuous-time state space models that accurately characterize the UAV attitude behavior. The identified models achieved a mean squared error (MSE) below 0.002, indicating strong agreement between predicted and measured responses. These validated models formed the basis for tuning a PID controller, significantly enhancing flight stability and responsiveness. The proposed framework, which combines rig design, precise data acquisition, and data-driven modeling, provides a practical and scalable solution for UAV developers to perform pre-flight validation, reduce tuning time, and improve operational safety and control performance.