摘要
针对目前卫星导航信号无法满足室内高定位精度需求的问题,首次结合室内场景的建筑特点,实地勘测了数据集,提出了一种改进行人航位推算(PDR)与接收信号强度指示(RSSI)融合的室内定位方法。针对RSSI易受环境干扰问题,通过采集蓝牙5.1信标数据,提出了一种基于RSSI的蓝牙定位卷积神经网络(CNN)方法,有效提高了定位精度;为解决PDR方法存在的定位误差累积问题,首次根据不同位置停留时间长短不同的特点,提出了一种基于停留时长的修正PDR(LOS-MPDR)算法;在MPDR定位和基于RSSI定位分析的基础上,使用扩展卡尔曼滤波(EKF)将2种方法融合,进一步提高定位精度。实验结果表明:本文方法的定位误差为0.28 m,满足应用场景需求。
Aiming at the problem that current satellite navigation signal can not meet the needs of indoor high positioning precision,combined with the architectural characteristics of indoor scenes,a set of datasets is investigated for the first time,and an indoor positioning method fuses improved pedestrian dead reckoning(PDR)and recieved signal strength indication(RSSI)is proposed.Aiming at the problem that RSSI is easily disturbed by environment,a Bluetooth positioning convolutional neural network(CNN)method based on RSSI is proposed by collecting Bluetooth 5.1 beacon data,which effectively improves the positioning precision.In order to solve the problem of positioning error accumulation in PDR method,a modified PDR correction algorithm based on length of stay(LOS-MPDR)in different position is proposed for the first time according to the characteristics of different dwell time at different locations.On the basis of MPDR positioning and RSSI-based positioning analysis,the extended Kalman filtering(EKF)technology is used to fuse the two methods to further improve the positioning precision.The experimental results show that the positioning error is 0.28 m,which can meet the needs of application scenarios.
作者
艾青
杨俊杰
蒋伟
隋志成
罗钦扬
张露明
AI Qing;YANG Junjie;JIANG Wei;SUI Zhicheng;LUO Qinyang;ZHANG Luming(School of Electronic and Information Engineering,Shanghai University of Electric Power,Shanghai 201306,China;School of Electronic Information,Shanghai DianJi University,Shanghai 201306,China;Shanghai Yancan Electronic Technology Co Ltd,Shanghai 201306,China)
出处
《传感器与微系统》
CSCD
北大核心
2023年第12期75-78,82,共5页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(61202369,61401269,61572311)
上海市技术创新项目(17020500900)
上海市教育发展基金会和上海市教委资助的“曙光计划”项目(17SG51)。
关键词
低功耗蓝牙室内定位
卷积神经网络
行人航位推算
扩展卡尔曼滤波
融合定位
low power consumption Bluetooth indoor positioning
convolutional neural network(CNN)
pedestrian dead reckoning(PDR)
extended Kalman filtering(EKF)
fusion positioning
