【作者】 苏丽； 李恩泽； 刘志林；

【Author】 Li Su;Enze Li;Zhilin Liu;College of Automation,Harbin Engineering University;

【机构】 哈尔滨工程大学自动化学院；

【摘要】 移动机器人的同步定位与地图构建(简称SLAM)问题是整个移动机器人自主导航的基础。论文将鼠脑海马结构认知机理应用在机器人SLAM问题中,基于啮齿类动物神经模型的RatSLAM算法来解决移动机器人即时定位与建图问题。有别于基于概率模型的SLAM方法,该方法将代表机器人位姿的位姿细胞与代表不同场景的局部场景单元相互关联,并利用不同位置的视觉信息的差异来进行定位并构建相应的拓扑地图。论文在室外环境中对RatSLAM进行了离线仿真实验,验证了算法的有效性;并在相同的室内环境下,对基于概率模型的EKF-SLAM算法和RatSLAM算法的实验结果进行了对比分析,实验结果表明,RatSLAM算法具有自我修正功能,在闭合环境下具有更好的运行结果。更多还原

【Abstract】 Simultaneous localization and mapping(SLAM) problem is the basis for autonomous navigation of the entire mobile robot.In this paper,the cognitive mechanism of rat hippocampus is applied to the SLAM problem of robot,and RatSLAM algorithm based on neural model of rodent is used to solve the problem of simultaneous localization and mapping of mobile robot.Different from the probabilistic model based SLAM method,this method correlates pose cells representing robot pose with local scene units representing different scenes,and uses the difference of visual information of different locations to locate and construct the corresponding topological map.In the outdoor environment,RatSLAM was simulated offline to verify the its effectiveness.The experimental results of EKF-SLAM algorithm and RatSLAM algorithm are compared under the same indoor environment.The experimental results show that RatSLAM algorithm has better performance in closed environment because of its self-correcting function.更多还原