UWB-VO: Ultra-Wideband Anchor Assisted Visual Odometry

This paper proposes a novel tightly coupled simultaneous localization and mapping (SLAM) method that combines monocular vision and ultra-wideband (UWB) technology. The proposed approach utilizes UWB distance information to restore the scale of monocular SLAM with an error rate of less than 1%. Our approach joint optimizes the distance residual between the camera and anchor, and the reprojection error of the map point, to improve SLAM positioning accuracy and mitigate visual SLAM degradation in challenging environments. A novel UWB anchor position optimization method is proposed, which enables the system to complete initialization within 2 seconds. The experimental results show that the proposed fusion scheme achieves significantly higher accuracy in system scale estimation and positioning compared to current state-of-the-art SLAM and visual-inertial odometry algorithms.