A Model-Based Wind Estimation Method for Unmanned Helicopter

Abstract

Atmospheric wind and turbulence significantly affect flight efficiency, control performance, and safety. This paper describes a new original in-flight algorithm for wind/turbulence field estimation using a standard unmanned rotorcraft sensor kit and control signals produced by the flight control system. A linear identified model is proposed as a basic model of helicopter dynamics. This model is modified into a quasi-nonlinear model by adding trim data and nonlinear kinematic equations. A nonlinear Unscented Kalman Filter is designed using the quasinonlinear model and the scaled unscented transformation to estimate wind components in the North-East-Down frame. The resulting estimate is split using moving averaging into the steady horizontal wind and atmospheric turbulence. This algorithm has been tested on a continuous model-stitched simulation in which the wind and atmospheric turbulence simulated by the Dryden model have been added. The result was also validated by an actual Steadicopter Black Eagle (BE-50) unmanned helicopter flight test.