Globally exponentially stable wind estimation for small fixed-wing UAVs using standard sensors

Abstract

This paper considers the estimation of the steady wind component around small fixedwing unmanned aerial vehicles using measurements from a Pitot-static tube and velocity over ground information. A three-stage filter is proposed, which yields globally exponentially stable estimation of the wind velocity and a correction factor on the dynamic pressure measurements from the Pitot-static tube. The suggested filter is compared with an extended Kalman filter and shown to yield identical steady state performance in both simulations and experiments. Additionally, the improved stability with respect to the extended Kalman filter is demonstrated with a choice of suboptimal tuning, for which the extended Kalman filter failed to converge whereas the suggested filter did not.