CUAHN-VIO: Content-and-uncertainty-aware homography network for visual-inertial odometry

Yingfu Xu, Guido C. H. E. de Croon

Abstract: Highlights•Our self-supervised uncertainty-aware planar homography network generalizes well.•Cascaded networks making use of EKF’s a priori state benefits efficiency.•CUAHN-VIO is the first learning-based VIO running onboard to navigate a flying robot.