Wind Speed / Direction Estimation of Quadrotor UAV Based on BP Neural Network

Kai-Hong Wu

Abstract: In view of the problem that the rotor UAV only uses theIMU to estimate the ambient wind speed / direction, the existing literature generally solves the problem through the UAV's dynamic model or aerodynamic model [1], but the models used are more simplified and complex. In this study, by establishing the function fitting relationship between the ambient wind speed / wind direction and the conventional airborne IMU and other relevant flight parameters of the rotorcraft, the estimation of the ambient wind speed / direction in the flight process of the rotorcraft can be completed.

Keywords: BP Neural Network, Quadrotor, Wind speed, direction estimation, Ambient air flow