未知环境下基于GNG网络环境地图自主认知.pdfVIP

Proceedings of the 29th Chinese Control Conference July 29-31, 2010, Beijing, China 未知环境下基于GNG 网络的环境地图自主认知* 仲朝亮，刘士荣, 邱雪娜 州电子科技大学 自动化研究所, 杭州 310018 E-mail: cpenet@163.com 摘 要：对环境地图的认知包含地图信息的获取和理解两个方面。以智能机器人自身对环境认知的要求，提出一种基 于Mental Model Imprinting (MMI)认知框架的环境地图认知模型。该模型利用Growing Neural Gas(GNG)算法的增长特 性，通过不断增加新的拓扑网络节点对周围环境的整体性知识进行抽象提取，构建易于机器人理解的环境地图。该认 知模型具有异步、自学习、自适应等特点。仿真结果验证其正确性与有效性。 关键词：环境认知, 拓扑网络, 移动机器人 Autonomous Cognitive of the Environment Map Based on * Growing Neural Gas in Unknown Environment ZHONG Chaoliang, LIU Shirong QIU Xuena （Institute of Automation, Hangzhou Dianzi University, Hangzhou 310018, P.R.China Abstract: Environmental map cognition is composed of two aspects of map information acquision and comprehension. From the intelligent robot s own demands for environmental cognition, a map cognitive model is proposed based on the Mental Model Imprinting (MMI) framework. Using the growth characteristics of Growing Neural Gas (GNG) algorithm, this model abstracts the holistic knowledge of the surrounding environment by adding new node of topology network and builds an environmental map which robot can understand easily. This cognitive model is asynchronous, self-learning and adaptive. Simulation results show the feasibility and effectiveness of the proposed method. Keywords: Environmental Cognition, Topological Network, Mobile Robot 移动机器人的空间环境特征。度量地图建模又分为栅 1 引言(Introduction) 格地图建模和几何地图建模。栅格地图易于表述环境 人们到达一个陌生的城市，在参观这个城市的同 的空间信息及不确定性，但它受限于环境的规模。几 时在脑海中轻松地构建了这个城市的地图，此后可以 何地图使用抽象的几何特征，例如点、线段或曲线来 轻松地识别身