Federated Filter Based on Navigation System of Quad-rotor UAV

In order to reduce the divergence of navigation results during GNSS outage for quad-rotor, and to solve the problem of collaborative work of multiple subsystems, we propose an integrated navigation system based on motion prediction and federated filtering. The proposed framework consists of a Micro Strap down Inertial Navigation System (MSINS), a Global Navigation Satellite System (GNSS), an Attitude and Heading Reference System (AHRS) and a motion prediction module. When GNSS signals are available, AHRS is used to get accurate attitude angle information, and information from MSINS, AHRS and GNSS is fused together by a federated filter to reduce navigation errors. When GNSS signals are unavailable, a motion prediction module based on the motion model of quad-rotor is introduced, and the predicted movements aid MSINS to restrain the divergence of position results. The system is tested on a self-built quad-rotor platform. Results show that the system can provide accurate attitude and location information during GNSS outage in short-term.