融合UWB的移动机器人视觉定位算法

he prerequisite for autonomous navigation of mobile robots is the accurate positioning of itself. Aiming at the problems that the visual simultaneous localization and map building (SLAM) algorithm has difficulty in eliminating the cumulative error of localization in large scenes and feature-similar environments, and the localization accuracy still needs to be improved, a visual localization algorithm for mobile robots that incorporates the ultra-wideband (UWB) technology is proposed. Firstly, the UWB tag node is localized using the maximum llikelihood estimate localization method; secondly, based on the nonlinear optimization method, the localization residuals function of positoning information from the UWB nodes are introduced into the visual SLAM tracking and localization module to optimize the estimation of the mobile robot position; finally, simulation experiments are carried out using the TUM public dataset. The simulation experiments show that the algorithm is superior to single UWB localization or vision SLAM localization approaches.