一种基于迭代EKF的FastSLAM算法

山东大学学报(工学版) ›› 2012, Vol. 42 ›› Issue (4): 41-47.

一种基于迭代EKF的FastSLAM算法

张丽,赵春霞*

南京理工大学计算机科学与技术学院, 江苏南京 210094)〖KH+6.5mmD

收稿日期:2012-05-06 出版日期:2012-08-20 发布日期:2012-05-06
通讯作者: 赵春霞(1964- )，女，北京人，教授，博士生导师，主要研究领域为智能机器人与智能检测系统，图形图像技术等. E-mail: zhaochunxia@126.com E-mail:zhaochunxia@126.com
作者简介:张丽(1988- )，女，安徽六安人，硕士研究生，主要研究方向为智能机器人系统导航研究. E-mail: yingzisashuang1988@126.com
基金资助:
高等学校博士点专项基金资助项目(20093219120025);国家自然科学基金资助项目(61101197)

A new FastSLAM algorithm based on iterated EKF

ZHANG Li, ZHAO Chun-xia*

College of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing 210094, China

Received:2012-05-06 Online:2012-08-20 Published:2012-05-06

摘要/Abstract

摘要：

针对在传统的快速地图创建和同时定位算法(fast simultaneous location and map building, FastSLAM)中采用扩展卡尔曼滤波器(extend Kalman filter, EKF)来估计机器人位姿和地图创建所带来的线性化误差的问题,本研究提出了一种基于迭代EKF的FastSLAM2.0算法——IFastSLAM算法。该算法将迭代思想运用到EKF中,同时采用迭代EKF来估计粒子从而完成机器人地图创建和自身定位。实验结果证明,该算法提高了粒子的估计精度从而减缓粒子退化问题,并更好的维持了地图的一致性。

关键词: FastSLAM2.0算法, 迭代EKF, IFastSLAM算法

Abstract:

The traditional fast map building and positioning algorithm for fast simultaneous location and map building (FastSLAM)usually used the extend Kalman filter (EKF)to estimate the robot’s pose and map, which could lead to some problems of linearization error. In order to solve this problem, a new FastSLAM2.0 algorithm based on the iterated EKF was proposed, which were also called IFastSLAM algorithm. The iterated EKF were used to estimate the particle and then to complete the map building and selfpositioning. The experimental results showed that this algorithm could improve the accuracy of estimating particle to slow down the particle degradation, and could maintain the consistency of the map better.

Key words: FastSLAM2.0 algorithm, the iterated EKF filter, IFastSLAM algorithm

张丽,赵春霞*. 一种基于迭代EKF的FastSLAM算法[J]. 山东大学学报(工学版), 2012, 42(4): 41-47.

ZHANG Li, ZHAO Chun-xia*. A new FastSLAM algorithm based on iterated EKF[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2012, 42(4): 41-47.

参考文献

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed