计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (36): 4-6.

• 博士论坛 • 上一篇    下一篇

一种改进的FastSLAM算法

康叶伟1,黄亚楼2,孙凤池2,苑 晶1   

  1. 1.南开大学 信息技术科学学院,天津 300071
    2.南开大学 软件学院,天津 300071
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-12-21 发布日期:2007-12-21
  • 通讯作者: 康叶伟

Improved FastSLAM algorithm

KANG Ye-wei1,HUANG Ya-lou2,SUN Feng-chi2,YUAN Jing1   

  1. 1.College of Information Technical Science,Nankai University,Tianjin 300071,China
    2.College of Software,Nankai University,Tianjin 300071,China

  • Received:1900-01-01 Revised:1900-01-01 Online:2007-12-21 Published:2007-12-21
  • Contact: KANG Ye-wei

摘要:

同时定位与建图是移动机器人实现真正自治的必要前提,FastSLAM作为一种成功的SLAM方法受到研究者的广泛青睐,FastSLAM将SLAM问题分为一个定位问题和一个建图问题,其中用扩展卡尔曼滤波器(EKF)实现地图陆标的估计与更新,提出了一种改进的FastSLAM方法,用UKF滤波器代替EKF实现FastSLAM中的陆标估计,使得陆标的估计精度提高,该方法同时具有UKF滤波器无需求解观测模型的雅克比矩阵的优点。

关键词: 同时定位与建图, UKF, 扩展卡尔曼滤波器, FastSLAM

Abstract: Simultaneous localization and mapping,for short SLAM,is a necessary prerequisite to make mobile robot truly autonomous,which is a hot research topic today.FastSLAM as a successful SLAM method abstracts many researchers’ attentions.FastSLAM factors the SLAM problem into a localization problem and a mapping problem in which the landmark position is estimated by EKF.A modified FastSLAM is presented,using the UKF to replace the EKF to estimate the landmark position.So we can improve the estimation precision,at the same time no need to linearize the sensor observation mode and to compute its Jacobian matrix.

Key words: Simultaneous Localization and Mapping(SLAM), Unscented Kalman Filter(UKF), Extended Kalman Filter(EKF), FastSLAM