Ultra-wideband and Visual Odometry Based Relative Localization for Multi-UAV System

This paper proposes a relative localization algorithm for Multi-UAV system, in which each agent is able to estimate the relative position of its neighbor based on Ultra-wideband measurement and visual odometry information exchanged. The relative localization algorithm consists of two parts. Visual odometry pre-process part fuses the IMU data and radar altimeter data to correct visual odometry error and reach a higher update rate to ensure flight smoothly and a better state estimation. Using the UWB measurement and information exchanged, each agent constructs a graph to solve relative localization problem in sliding window. Finally, experimental results are presented to corroborate the effectiveness of our proposed relative localization algorithm.