A Practical Method for Quadrotor Modeling and Parameter Identification

This article introduces a methodology for modeling and parameter identification of the quadrotor systems, which obviates the need for external devices. The model parameters and inherent system disturbances can be identified via flight data acquisition through closed-loop system flight testing of an unmanned aerial vehicle (UAV). This study initially simplifies the model for each axis into second-order systems with disturbances, which necessitates the identification of only three characteristic parameters, thereby alleviating the complexity of the identification process. Subsequently, the identification of optimal model parameters is achieved via closed-loop measurement data. The subsequent physical experiment validates that the accurancy of Normalized Mean Square Error (NMSE) for the three-axis angles and angular velocities exceeds 90%. The proposed algorithm constructs an inner-loop controlled object model for quadrotor systems, facilitating the design of closed-loop control system for both angular velocity and outer loops.