An aerodynamic model identification method suitable for low-quality flight data

Aiming at the problem of aerodynamic model identification of multiple low-quality flight test data, a modified multivariate orthogonal functions method and its application framework are proposed in this paper. Based on the traditional multivariate orthogonal functions, this modified approach firstly introduces the total least square to solve the influence of flight data measurement noise on parameter estimation. Then, according to the model stingy modeling criterion, the iterative idea and t statistic metric are applied to make full use of the flight data lacking effective excitation. The simulation results show that the modified multivariate orthogonal functions approach can identify more compact and accurate model structure and aerodynamic parameters than the traditional method. The flight test results show that the aerodynamic correction model and uncertainty model based on the identification can effectively implement the correction of the ground aerodynamic database and improve the accuracy of the aerodynamic characteristics prediction, thus laying the foundation for the simulation evaluation and controller design of the unmanned aerial vehicles.