摘要
近年来,基于Wi-Fi接收信号强度的室内定位技术一直是研究领域的热点问题.随着智能家居和可穿戴计算的高速发展,大量新型智能设备的出现进一步推动了室内定位技术的发展,同时也带来了新的挑战.可穿戴设备与传统智能设备相比,有着与用户更加紧密的位置绑定关系,是一类更加适合的室内定位平台.但另一方面,类似智能手表、眼镜、手环以及戒指等可穿戴设备,由于其自身资源受限的特性,迫切需要一种低功耗的新型室内定位算法.通过本文在Moto 360二代智能手表上进行的实验可以发现,基于Wi-Fi RSS的室内定位服务会使智能手表的使用时间缩短82%以上,其中99%的定位能耗,是为了保证定位精度而大量进行射频信号采集所造成的.简单的减少信号采集量将带来显著的定位精度下降,如何在保障定位精度的前提下,尽可能地减少信号采集量是低功耗定位技术面临的核心挑战.该文提出了一种面向能耗受限的可穿戴设备的室内定位技术TinyLoc.TinyLoc在实时定位阶段仅需要一次信号采集,同时运用用户运动特性弥补信号采集量减少而带来的精度缺失.实验结果表明,在90%的情况下,TinyLoc对于完整路径上的点平均误差可以达到3m以内,另一方面,在相同实验环境下,TinyLoc能耗为传统Wi-Fi定位算法的1/6,是MoLoc的64%.相比传统的基于Wi-Fi信号的定位算法,TinyLoc可以延长Moto 360二代智能手表约3倍的定位工作时间.
Indoor localization technology based on Wi-Fi has long been a hot research topic in the past decade.Despite of numerous solutions,new challenges have been arisen along with the trend of smart home and wearable computing.Compared with the traditional intelligent devices,the wearable devices have a tighter position binding relationship with users,and are a more suitable indoor localization platform.However,power efficiency needs to be significantly improved for resource-constrained wearable devices,such as smart watch,wristband,finger ring,etc.According to our experiments on 2nd generation MOTO 360 smart watch,indoor locating service will reduce the standby time of the device by 82%,and 99% of the system energy consumption can be attributed to real-time radio scan.However,simply reducing radio data collection will cause a serious loss ofpositioning precision because of unstable Wi-Fi signals.Reducing the amount of data collection with ensuring locating accuracy is the core challenge of energy-constrained indoor localization.In this paper,we present TinyLoc,an indoor localization approach that only needs one real-time radio scan in the localization phase to achieve energy-saving performance for wearable devices.Meanwhile,TinyLoc enhances the localization accuracy using users＇ motion features.Experiment results demonstrate that location point error is less than 2 meters for more than 90% cases on the full path,and energy consumption is only 1/6 of the traditional Wi-Fi localization algorithm and 64% of MoLoc in the same experimental settings.Compared to traditional locating algorithms based on Wi-Fi signals,TinyLoc can extend about three times of the standby time for indoor locating on 2nd generation Moto 360 smart watch.
出处
《计算机学报》
EI
CSCD
北大核心
2017年第8期1813-1828,共16页
Chinese Journal of Computers
基金
国家自然科学基金(61170292
61472212)
国家科技重大专项课题(2015ZX03003004)
国家"八六三"高技术研究发展计划项目基金(2013AA013302
2015AA015601)
欧盟CROWN基金项目(FP7-PEOPLE-2013-IRSES-610524)资助~~
关键词
室内定位
节能
运动特性
可穿戴计算
传感器网络
物联网
信息物理融合系统
indoor localization
energy efficiency
user motion
wearable computing
sensor networks
Internet of Things
Cyber-Physical System
作者简介
王晓亮，男，1986年生，博士研究生，主要研究方向为无线网络和移动计算．E-mail：wangxiaoliang12@mails．tsinghua．edu．cn．
徐恪，男，1974年生，博士，教授，博士生导师，主要研究领域为互联网体系架构、高性能路由器、P2P网络、物联网和网络经济学．
杨铮，男，1983年生，博士，副教授，主要研究方向为无线网络与移动计算．
葛志诚，男，1992年生，硕士研究生，主要研究方向为物联网和传感器网络．
