TY - GEN

T1 - TagTrack

T2 - 11th International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services, MobiQuitous 2014

AU - Ruan, Wenjie

AU - Yao, Lina

AU - Sheng, Quan Z.

AU - Falkner, Nickolas J.G.

AU - Li, Xue

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Device-free passive localization aims to localize or track targets without requiring them to carry any devices or to be actively involved with the localization process. This technique has received much attention recently in a wide range of applications including elderly people surveillance, intruder detection, and indoor navigation. In this paper, we propose a novel localization and tracking system based on the Received Signal Strength field formed by a set of cost-efficient passive RFID tags. We firstly formulate localization as a classification task, where we compare several state-of-theart learning-based classification methods including k Nearest Neighbor (kNN), Multivariate Gaussian Mixture Model (GMM) and Support Vector Machine (SVM). To track a moving subject, we propose two HiddenMarkovModel (HMM)- based methods, namely GMM-based HMM and kNNbased HMM. kNN-based HMM extends kNN into a probabilistic style to approximate the Emission Probability Matrix in HMM. The proposed methods can be easily applied into other fingerprint-based tracking systems regardless of their hardware platforms. We conduct extensive experiments and the results demonstrate the effectiveness and accuracy of our approaches with up to 98% localization accuracy and an average of 0.7m tracking error.

AB - Device-free passive localization aims to localize or track targets without requiring them to carry any devices or to be actively involved with the localization process. This technique has received much attention recently in a wide range of applications including elderly people surveillance, intruder detection, and indoor navigation. In this paper, we propose a novel localization and tracking system based on the Received Signal Strength field formed by a set of cost-efficient passive RFID tags. We firstly formulate localization as a classification task, where we compare several state-of-theart learning-based classification methods including k Nearest Neighbor (kNN), Multivariate Gaussian Mixture Model (GMM) and Support Vector Machine (SVM). To track a moving subject, we propose two HiddenMarkovModel (HMM)- based methods, namely GMM-based HMM and kNNbased HMM. kNN-based HMM extends kNN into a probabilistic style to approximate the Emission Probability Matrix in HMM. The proposed methods can be easily applied into other fingerprint-based tracking systems regardless of their hardware platforms. We conduct extensive experiments and the results demonstrate the effectiveness and accuracy of our approaches with up to 98% localization accuracy and an average of 0.7m tracking error.

KW - Gaussian mixture model

KW - Hidden markov model

KW - Kernel-based

KW - Localization

KW - Nearest neighbor

KW - Rfid

U2 - 10.4108/icst.mobiquitous.2014.258004

DO - 10.4108/icst.mobiquitous.2014.258004

M3 - Conference contribution/Paper

AN - SCOPUS:84924311143

SP - 80

EP - 89

BT - MobiQuitous 2014 - 11th International Conference on Mobile and Ubiquitous Systems

PB - ICST (Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering)

Y2 - 2 December 2014 through 5 December 2014

ER -