System Identification for the HA-420 HondaJet: A Data-Driven Approach to Aerodynamic Parameter Estimation

The aerodynamic characterization of an aircraft is essential for optimizing performance, efficiency, and safety. Traditional methods for determining aerodynamic parameters—such as lift coefficient (CL), drag coefficient (CD), rolling moment coefficient (Cl), pitching moment coefficient (Cm), and yawing moment coefficient (Cn)—rely on extensive wind tunnel and flight testing, which are costly and time-intensive. System identification is the process of building mathematical models of dynamic systems based on observed data. It offers a computational alternative by deriving aerodynamic models from flight test sensor data, potentially reducing the number of required test flights and associated costs. This study develops and validates a novel system identification model for estimating the HA-420 HondaJet’s aerodynamic coefficients using recorded flight test data. A linear system identification method was developed, employing ordinary least squares (OLS) regression to minimize the cost function between the estimated and theoretical kinematic aerodynamic coefficient models. Computational analysis was conducted in MATLAB, with results validated against wind tunnel data from Honda Aircraft Company. Findings indicate that the system identification approach accurately predicts aerodynamic coefficients when the flight data includes perturbations sufficient to excite the aircraft’s dynamic modes. However, when applied to data with limited variations in pitch, roll, and yaw, the model failed to capture key aircraft characteristics, leading to significant deviations from expected aerodynamic behavior. These results provide insight into the minimum perturbation requirements for reliable system identification modeling. Furthermore, they highlight the potential of system identification to reduce reliance on costly experimental testing, improve model-based flight dynamics analysis, and enhance pilot safety during flight.