Kalman Filters for Air Data System Bias Correction for a Fixed-Wing UAV

Abstract

This paper presents two Kalman filter approaches for correcting air data systems providing relative velocity measurements with an additive constant or slowly time-varying bias for fixed-wing unmanned aerial vehicles (UAVs). In addition to the air data system, both estimators rely on a standard sensor suite consisting of a GNSS receiver, an IMU, and a heading reference. Furthermore, the estimators are based on kinematics and do not require a model of the UAV. The two estimators are in the noise-free case proven to have globally exponentially stable (GES) equilibrium points of the error dynamics if provided with persistence of excitation (PE) of the angular velocity of the UAV. The estimators are verified through simulation and using experimental flight data. The relative velocity measurements in the experimental flight results are provided by an array of pressure sensors embedded in the surface of the UAV combined with a neural network algorithm. The results indicate that a certain amount of PE is needed in order to have converging bias estimates for turbulent wind conditions.