基于多分辨率地图位姿跳变优化的改进Cartographer算法  

Improved Cartographer Algorithm Based on Pose Jump Optimization of Multi-resolution Map

作　　者：陈斌斌 盛志超 叶林 杨强强 方勇[1] CHEN Binbin;SHENG Zhichao;YE Lin;YANG Qiangqiang;FANG Yong

机构地区：[1]上海大学通信与信息工程学院 [2]上海映驰科技有限公司

出　　处：《计量与测试技术》2024年第2期41-44,共4页Metrology & Measurement Technique

基　　金：国家自然科学基金(项目编号:61901254)。

摘　　要：针对Cartographer算法在局部扫描匹配中,存在匹配精度低、位姿跳变影响回环检测的问题,提出一种基于多分辨率地图位姿跳变优化的改进Cartographer算法。首先,在局部优化中,建立多分辨率栅格地图,提高初始位姿的估计精度;然后,在回环检测和全局优化之间,建立位姿跳变检测机制,通过减少突变节点数量,提高回环检测的可靠性;最后,通过数据集中,对改进算法进行验证。结果表明:该算法能有效提高定位精度和建图效果。To solve the problems of Cartographer algorithm in local scan matching,such as low matching accuracy and pose hop affecting loop detection,an improved Cartographer algorithm based on multi-resolution map pose hop optimization was proposed.Firstly,in local optimization,a multi-resolution raster map is established to improve the estimation accuracy of the initial pose.Then,between loop detection and global optimization,the pose jump detection mechanism is established to improve the reliability of loop detection by reducing the number of mutation nodes.Finally,through the data set,the improved algorithm is verified.The results show that this algorithm can effectively improve the positioning accuracy and the effect of map construction.

关 键 词：同步定位与地图构建 Cartographer算法 点云匹配 回环检测 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] TP391.41[自动化与计算机技术—控制科学与工程]