Nonlinear attitude and gyroscope's bias estimation for a VTOL UAV

This paper addresses the problem of attitude and heading restitution for a VTOL UAV (Vertical Take Off and Landing Unmanned Aerial Vehicle). In a first step, we propose an observation strategy to restore the attitude starting from a measure of the orientation matrix and angular rate readings polluted by an unknown bias vector. In the second step, we detail the implementation of this algorithm in the HoverEye, a ducted fan VTOL UAV developed by Bertin Technologies, equipped with an inertial measurement unit (IMU) and magnetometers. A measured orientation matrix is calculated from the gravity and the earth's magnetic field. Then, an estimated orientation is built by integration of gyroscopic readings, and corrected by the measured ones. The proposed observer, directly designed in the special orthogonal group SO(3), is as efficient as classical Extended Kalman Filtering based observers, but easier to implement in real time. Experiments achieved on the HoverEye illustrate the concept.