Data fusion and switching function for MAV navigation system

In this work, data fusion and switching function based approach for MAV (Micro Aerial Vehicle) navigation system is proposed. Our approach applies monocular simultaneous localization and mapping (SLAM) system. Extended Kaiman Filter (EKF) algorithm is then used to fuse SLAM data with velocity and acceleration data, yielding metric-scaled position data. Moreover, GPS sensor is employed to retrieve the MAV position in global world frame coordinate. In addition, the switching function is added to deal with particular conditions. Some benefits from our approach are improvement of data accuracy when the MAV is moving indoor with global coordinate system as well as overcoming the problem of losing visual tracking.