【摘要】 提出一种基于ORB(oriented FAST and rotated BRIEF)特征和里程计信息的移动机器人视觉导航方法。该方法提取输入场景的ORB特征作为显著路标,同时提取机器人里程计信息计算行为模式,将显著路标和行为模式作为状态节点进行机器人环境拓扑地图的构建。基于特征匹配技术进行机器人的状态定位,并选择状态中对应的行为模式实现机器人的路径跟踪导航,且在导航过程中有防止偏移既定路线的调整机制。实验表明:文中的方法能够较好的完成机器人的导航任务,并且对环境中的动态信息具有一定的鲁棒性。更多还原

【Abstract】 A mobile robot vision navigation method was proposed based on ORB( oriented FAST and rotated BRIEF) features and odometer information. The ORB features was extracted from input scenes as salient landmarks,meanwhile robot’s behavior was calculated based on odometer information. Both the landmarks and the behavior patterns was informed as a state to build the topological map of the environment. Robot localization was processed by feature matching,and behavior pattern was selected in corresponding state to realize robot’s path tracking and navigation,an adjustment mechanism was implemented to avoid the deviation from the path. Experimental results showthat: the method is able to perform good tracking and navigation task,and has robustness for dynamic information of environment.更多还原