仿鼠脑海马的机器人地图构建与路径规划方法  

Robotic map building and path planning methods based on the hippocampus of rat’s brain

作　　者：邹强 丛明[1,2] 刘冬 杜宇[3] Zou Qiang;Cong Ming;Liu Dong;Du Yu(School of Mechanical Engineering,Dalian University of Technology,Dalian 116024,Liaoning China;Dalian University of Technology Jiangsu Research Institute Co.Ltd.,Changzhou 213164,Jiangsu China;Dalian Dahuazhongtian Technology Co.Ltd.,Dalian 116023,Liaoning China)

机构地区：[1]大连理工大学机械工程学院,辽宁大连116024 [2]大连理工江苏研究院有限公司,江苏常州213164 [3]大连大华中天科技有限公司,辽宁大连116023

出　　处：《华中科技大学学报（自然科学版）》2018年第12期83-88,共6页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61503057);辽宁省自然科学基金联合基金资助项目(20180520017);江苏省智能装备产业技术创新中心产业技术联合研发资金资助项目

摘　　要：针对移动机器人在非结构化环境下的导航任务,根据哺乳动物海马体空间细胞的认知机理,提出了一种仿鼠脑海马的机器人情景认知地图构建及路径规划方法.在机器人情景记忆建模过程中集成位置细胞与网格细胞神经元活动机制,建立机器人空间环境情景认知地图,采取状态神经元集合序列全局路径规划策略,在记忆空间以自我为参考,通过事件再配置预测并规划最优情景轨迹.实验结果表明:该方法能够生成精确的情景认知地图,并且基于目标导航能够规划一条最佳路径.Inspired by the biological cognitive mechanism of hippocampal spatial cells in mammals,an episodic cognitive map building and global path planning methods were proposed for the mobile robotic navigation tasks in an unstructured environment.The activity mechanism of place cells and grid cells were integrated into the episodic memory modelling,and an episodic cognitive map was built for the robotic spatial environment.The proposed global path planning strategy was adopted based on state neurons sequence reorganization.The mobile robot could localize itself relative to its past experiences in the memory space,and then anticipated and planned the future events sequence.The experimental results show that the proposed methods can build an accurate episodic-cognitive map and plan a preferred trajectory for robotic task navigation.

关 键 词：移动机器人 海马体 位置细胞 网格细胞 情景认知地图 路径规划 状态神经元 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]