Attitude Determination of Highly Dynamic Fixed-wing UAVs with GPS/MEMS-AHRS integration

Abstract

Taking advantage of the recent Micro-Electro-Mechanical Systems (MEMS), the global cost of the Unmanned Aerial Vehicles (UAVs) has been reduced. However, this reduction in size, power and price of the sensors comes at the expense of an increase in accuracy degradation making it more difficult to estimate the attitude of highly dynamic UAVs. Developing an efficient Attitude and Heading Reference System (AHRS) is then imperative where the integration of the Global Positioning System (GPS) and Inertial Navigation System (INS) can provide a more reliable and accurate AHRS. In this article, the development of a GPS/MEMS-INS system specifically designed for attitude determination of fixed-wing UAVs is attempted and its performance evaluated. An Extended Kalman Filter (EKF) is developed where the measurements equations are analytically solved in order to avoid the derivation of the Jacobian matrix. The algorithm makes use of GPS-derived accelerations. Simulation results show that the attitude of the UAV can be accurately estimated, with maximum error standard deviations rounding one degree. The EKF algorithm was also tested with real flight data and results show a consistent roll, pitch and yaw angles estimation. Comparisons were made with a commercial device (MTi-G Xsens) and the innovation sequences of the EKF algorithm support its reliability.