视觉导航有穷自动机在自然语言状态中的应用  

Application of finite automatic machine based on vision navigation for state transition of natural language

作　　者：杨晶东[1] 杨敬辉[2] 

机构地区：[1]上海理工大学光电信息与计算机工程学院 [2]上海第二工业大学经济管理学院

出　　处：《黑龙江大学自然科学学报》2013年第2期276-280,共5页Journal of Natural Science of Heilongjiang University

基　　金：国家自然科学基金资助项目(60874002);上海市教委科研创新项目(B51JG13R003)

摘　　要：室内移动机器人自主导航是一系列自然语言状态与动作转换结果,有效地控制和描述机器人状态和动作转换是自主导航过程的关键。本文首先利用视觉传感器获取障碍物YUV色彩空间信息,采用区域增长法识别障碍物连通域,将机器人状态以及控制指令转化为抽象的形式符号并建立状态转移函数,进而建立基于视觉导航机器人有穷自动机模型。实验结果表明,该方法能平滑、准确地完成自然语言状态下的机器人自主导航过程,具有较高的实时性和鲁棒性。The autonomous navigation for mobile robots indoor is the results of some series state or action transition for natural language. It is the key how to control and describ the state or action under the control of natural language in effect. Firstly, the YUV color space of the object is acquired by camera sensor, region growing method is used to recognize the connected domain. The states of robot and control command are transformed to some abstract symbols and the state transition functions. Further, the finite automatic machine（FAM） model has been built based on vision navigation for mobile robot finally. The experiment results show that the method can finish the transition between the actions and states of mobile robots under natural language state efficiently and smoothly. It provides higher real-time and robustness for navigation

关 键 词：自然语言 有穷自动机 状态转移函数 移动机器人 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]