移动机器人同时定位和地图创建的一种新方法

为了实现室内环境中移动机器人的同时定位和地图创建，提出了一种新方法，该方法应用顺序蒙特卡罗方法，把定位和地图生成分成状态估计和参数估计，应用多粒子滤波器实现同时估计机器人位置和障碍物位置。环境感知采用机器人自身配置的声纳传感器，应用哈夫变换提取环境障碍物边界特征，用直线近似表示，并提出了一种可靠的特征匹配方法。Pioneer 2移动机器人在实际室内环境中完成的实验证明了该方法的可行性。

Novel Method of Mobile Robot Simultaneous Localization and Mapping

This paper provides a novel method to realize the mobile robot indoor simultaneous localization and mapping(SLAM),which is a key prerequisite for a truly autonomous robot.This method applies Sequential Monte Carlo(SMC) method which has made much success in localization recently.The localization and mapping is divided as state estimation and parameter estimation respectively,and multiple particle filters are applied to estimate the robot position and the obstacle position simultaneously.The environment information is observed through sonar sensors mounted on a robot itself.The Hough Transform is used for extracting environment obstacle limit feature from the sonar observation.The feature is represented by line approximately,and the robust algorithm is proposed to implement feature matching.The experiment is performed with a Pioneer 2 mobile robot in a real-world indoor environment,and the feasibility of this method is proved.