中文摘要：

提出了一种基于Hough变换的移动机器人全局定位方法。在机器人执行全局定位的初始阶段，将给定的环境栅格地图通过离散Hough变换转换为Hough空间中形成参考地图，并建立全局地图的Hough能量谱函数：而机器人在定位过程中通过激光传感器数据实时构建局部地图，并转换到Hough空间建立局部地图的Hough能量谱函数；构建全局和局部地图的Hough能量谱相关函数，通过局部最大化该函数求出两个地图问的旋转角度集。基于求得的角度集，对局部地图进行旋转配准，与全局地图一起构建Hough变换相关函数，通过局部最大化该函数可求出两个地图间的平移集合；应用求得的旋转角度集和平移集合，很容易估计机器人在全局地图中的位姿分布，通过不连续的几个时刻的位姿更新最终能够实现机器人全局定位。实验结果证明了该方法的有效性。

英文摘要：

A Hough transform based on the global positioning method of mobile robot is proposed in the paper. In execution of the initial phase of robot global localization, the given grid map of an environment is transformed into the reference map in Hough space by discrete Hough transform, and a Hough energy spectrum function is established for global map. In robot positioning process, the robot builds in real-time a local map through its laser sensor data and con- verts this map to Hough space to establish a Hough energy spectrum function. Subsequently, Hough energy spectrum cor- relation function of global and local map is constructed. Through the local maximization of energy related function, a rota- tion angle set can be worked out between the two maps. Based on the obtained set, the local map is rotated and applied to build Hough transform correlation function with global map. Through the local maximization of the function, the transla- tion set of two maps can be solved. Based on rotation angle set and translation set, the robot in the global map can be es- timated. By several discontinuous poses the robot can update its states and realize the global positioning. The experimental results show the effectiveness of the method.