Paper Info
Title
Robust fingerprinting-based localization using directed graphical models
Abstract
In this paper, we propose a robust fingerprinting-based localization using directed graphical model. To overcome the influence caused by the jitter of received signal strength (RSS), the location of one mobile node can be estimated by fusing both the current matching result and the previous location estimation, using the Bayesian graphical model (BGM). Then, the localization problem is cast as a maximum a posteriori (MAP) estimator, which is also proved to coincide to maximum likelihood (ML) estimator. However, the initialized MAP estimator can be hardly solved with incomplete statistical property concerning the random vectors. To this end, we propose an adaptive smoothing algorithm (ASA) to attain the suboptimal solution of the original problem. Finally, the experimental results show that the proposed algorithm obtains a significant performance gain.
Year
DOI
Venue
2017
10.1109/ICCChina.2017.8330332
2017 IEEE/CIC International Conference on Communications in China (ICCC)
Keywords
Field
DocType
Bayesian Graphical model (BGM),fingerprint matching,maximum a posteriori (MAP) estimator,adaptive smoothing algorithm (ASA)
Computer science,Algorithm,Real-time computing,Smoothing,Jitter,Graphical model,Maximum a posteriori estimation,Hidden Markov model,RSS,Bayesian probability,Estimator
Conference
ISBN
Citations 
PageRank 
978-1-5386-4503-1
0
0.34
References 
Authors
11
6
Name
Order
Citations
PageRank
Yueyue Zhang1207.77
Yaping Zhu2197.77
Weiwei Xia32814.30
Feng Yan47313.36
Lianfeng Shen551765.25
Yi Wu68518.02