摘要
由于移动机器人处在未知且不确定的环境中,主要采用基于概率的方法对同时定位与地图构建(SLAM)进行描述。建立了SLAM问题的概率表示模型,并对在解决SLAM问题中最常用的扩展卡尔曼滤波(EKF)算法以及迭代扩展卡尔曼滤波(IEKF)算法进行了描述。针对两种算法的缺陷和不足,将应用于跟踪领域的修正迭代扩展卡尔曼滤波算法(MIEKF)与SLAM思想结合,提出了一种新的基于MIEKF的SLAM算法,该算法能减小线性化误差并且不需要很高的观测精度。最后用上述算法进行了基于点特征的SLAM实验,验证了该算法的有效性。
As the mobile robot in the unknown and uncertain environment,the simultaneous localization and mapping(SLAM) mainly describled by probability.In this paper,established a probability model for the SLAM,and describe extended Kalman filter(EKF) algorithm which most commonly used in SLAM and iterated extended Kalman filter(IEKF) algorithm.Aiming the defects and deficiencies of these two algorithms,after combining the iterated extended Kalman filter algorithm(MIEKF) which applied to tracking field with SLAM thought,applied a novel SLAM algorithm based on MIEKF,this algorithm could reduce the linearization error and didn't require high observation precision.Finally,it also performed the point-feature-based SLAM experiments,and proved this algorithm works effectively.
出处
《计算机应用研究》
CSCD
北大核心
2011年第3期902-904,共3页
Application Research of Computers
基金
国际科技合作计划资助项目(2010DFA12160)
关键词
移动机器人
扩展卡尔曼滤波
修正
同时定位与地图构建
mobile robot
extended Kalnan filter(EKF)
modified
simultaneous localization and mapping(SLAM)