一种基于快速鲁棒特征匹配算法的类脑SLAM导航  

A brain-inspired SLAM navigation based on speeded up robust features matching algorithm

作　　者：王文照 唐军 申冲[1,2,3] 刘俊 WANG Wenzhao;TANG Jun;SHEN Chong;LIU Jun(School of Instrument and Electronics,North University of China,Taiyuan 030051,China;National Key Laboratory of Dynamic Testing Technology,North University of China,Taiyuan 030051,China;Shanxi Province Key Laboratory of Quantum Sensing and Precision Measurement,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学仪器与电子学院,山西太原030051 [2]中北大学动态测试技术国家重点实验室,山西太原030051 [3]中北大学山西省量子传感与精密测量重点实验室,山西太原030051

出　　处：《Journal of Measurement Science and Instrumentation》2023年第2期137-147,共11页测试科学与仪器（英文版）

基　　金：National Natural Sciences Foundation of China(Nos.61973281,51821003,51922009);Key Research and Development Project of Shanxi Province(No.202003D111003);Excellent Youngth Foundation of Shanxi Province(No.202103021222011);Foundation of Science and Technology on Electro-Optical Information Security control Laboratory(No.2021JCJQLB055010);Aviation Science Foundation(No.2018ZCU0002);Shanxi Province Key Laboratory of Quantum Sensing and Precision Measurement(No.201905D121001);1331 Project of Shanxi Province。

摘　　要：类脑导航是模拟鼠类感知环境机制提出的一种同步定位与构图(Simultaneous localization and mapping,SLAM)的导航算法。针对复杂环境如室内光线变化导致类脑SLAM导航产生误差的问题,本文提出了基于特征匹配(Speeded up robust features,SURF)算法的优化类脑SLAM导航模型。该模型通过一套移动视觉系统采集环境信息,构建的局部场景细胞通过SURF特征匹配算法获取到载体在环境中的方向与位置信息;头朝向细胞与位置细胞通过连续吸引子神经网络共同表示载体当前的位姿。利用所获取的位姿与时间信息,通过路径积分计算当前载体在坐标系中所处的位置;最后,构建基于认知点的拓扑经验地图。此外,在局部场景细胞获取环境信息的同时,通过SURF特征匹配算法来进行闭环检测,判断是否需要对当前位置进行修正。本文提出的优化类脑SLAM模型很大程度改进了原有模型在有光线变化的室内情况下易产生场景误匹配的问题,并通过实验验证了本文提出方法的有效性。Brain-inspired navigation algorithm controlled by simultaneous localization and mapping(SLAM)is prone to errors,primarily caused by complex environmental factors such as changes in light direction.To address this limitation,a braininspired SLAM approach is proposed where feature matching is supported by the speeded up robust features(SURF)algorithm.This model gathers environmental information via a set of mobile vision systems:the local view cell operators,designed to retrieve information about carrier direction/location via the SURF algorithm,and the head-direction cell and pose cell operators,which simultaneously represents current carrier position via continuous attractor neural networks.The position and time information retrieved by these cell operators are used to compute the position of the current carrier through path integration.In the final step,experience maps of the topology are constructed based on cognitive points.In addition,while the local view cell acquire environmental information,closed-loop detections are executed by the SURF algorithm to correct(if necessary)the current position.The optimized brain-inspired SLAM model successfully addresses the problem of scene matching errors faced by previous models in the presence of changing light direction.The effectiveness of the proposed method is validated through experimental verification.

关 键 词：类脑导航 同步定位和构图 特征匹配算法 经验地图 闭环检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]