融合UWB的移动机器人视觉定位算法
首发时间:2024-04-08
摘要:移动机器人自主导航的前提是对自身的精确定位。针对视觉同时定位与建图(SLAM)算法在大场景下和特征相似环境中定位累积误差难以消除,定位精度仍待提高等问题,提出一种融合超宽带(UWB)技术的移动机器人视觉定位算法。首先,使用极大似然估计定位法对UWB标签节点进行定位;其次以非线性优化法为基础,将融合UWB节点定位信息的残差函数引入视觉SLAM跟踪定位模块,优化移动机器人位姿估计;最后使用TUM公开数据集进行仿真实验。仿真实验表明,该算法优于单一UWB定位或视觉SLAM定位方式。
For information in English, please click here
A visual localization algorithm for mobile robots incorporating UWB
Abstract:The 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.
Keywords: Fusion Localization UWB vSLAM
基金:
引用

No.****
同行评议
勘误表
融合UWB的移动机器人视觉定位算法
评论
全部评论