一种新的移动机器人Monte Carlo自主定位算法

被引:7
作者
房芳
马旭东
戴先中
机构
[1] 东南大学自动化学院
基金
国家高技术研究发展计划(863计划);
关键词
移动机器人; Monte Carlo算法; 重要性函数; 过收敛检验; 均匀性检验;
D O I
暂无
中图分类号
TP242 [机器人];
学科分类号
1111 ;
摘要
针对当出现一些未建模的机器人运动时(如碰撞或者绑架问题),以小采样数目实现常规Monte Carlo方法难以解决的问题,提出一种新的Monte Carlo定位算法,该算法同时采用p(Xkzk)与p(XkXk-1)作为重要性函数并从中进行采样,避免了采样集不包含真实位姿采样的情况,能够有效地解决全局定位与绑架问题.同时在重采样过程中引入了过收敛检验与均匀性检验用于判断采样与感知信息的匹配程度,以适时进行重采样,节省了计算资源并提高了定位效率.实验结果表明该方法具有良好的性能.
引用
收藏
页码:40 / 44
页数:5
相关论文
共 10 条
[1]  
KLD-sampling:adaptive particle filters and mo-bile robot localization. Fox D. Technical Report,TR-UW-CSE-01-08-02 . 2001
[2]  
Mobile robot localization using a single im-age. Krotkov E. IEEE International Conference on Roboticsand Automation . 1989
[3]  
Sequential MCLmethods in practice particle:filters for mobile robot lo-calization. Fox D,Thrun S,Dellaert F,et al. . 2001
[4]  
Probabilistic robot navigation inpartially observable environments. Simmons R,Koenig S. InternationalJoint Conference on Artificial Intelligence . 1995
[5]  
Mobile robot localiza-tion by tracking geometric beacons. Leonard J,Durrant-Whyte H F. IEEE Transac-tions on Robotics and Automation . 1991
[6]  
A dapting the sample size in particle filterthrough KLD-Sampling. Fox D. International Journal ofRobotics Research . 2003
[7]  
Position verification of amobile robot using standard pattern. Kabuka M R,Arenas A E. IEEE Journalof Robotics and Automation . 1987
[8]  
MCL localization:effi-cient position estimation for mobile robots. Fox D,Burgard W,Dellaert F. Pro-ceedings of the AAAI-99 . 1999
[9]  
Bayesian filteringfor location estimation. Fox D,Hightower J,Liao J,et al. IEEE Pervasive Computing . 2003
[10]  
Feature based con-densation for mobile robot localization. Patric J,David A,Olle W,et al. IEEE In-ternational Conference on Robotics and Automation . 2000