臺灣博碩士論文加值系統

本論文提出了一種新穎的Wi-Fi室內定位系統架構，可以根據干擾自動調整切換深度學習系統的微調算法。該系統由離線階段和在線階段組成，於離線階段設計具有七個隱藏層的深層網路架構，利用受限玻爾茲曼機器的兩個連續隱藏層來堆疊而成，並使用對比度散度算法進行逐層訓練。本論文所提出的Wi-Fi室內定位方案使用自動調整微調算法。當環境中收集的訊號標準偏差不超過臨界值時，採用交叉熵算法來更新深層網路架構的參數;當環境中收集的訊號標準偏差超過臨界值時，採用均方算法來更新深度網路架構的參數。對於在線階段進行標準差判斷來選擇適合的指紋資料庫，再使用貝葉斯概率算法進行位置估計。實驗在干擾和無干擾的室內環境中實施和驗證，實驗結果表明，該方案可以提高訓練數據的準確性，降低定位誤差。

A novel Wi-Fi indoor localization system architecture, which can adaptively switch the fine-tuning algorithm of the deep learning system according to the interferences, is proposed in this paper. The system consists of the off-line and the on-line phases. A deep network architecture with seven hidden layers is designed for the off-line phase. Two continuous hidden layers from a Restricted Boltzmann Machine, which uses the k-step contrastive divergence algorithm for the layer-by-layer training. The proposed Wi-Fi indoor localization scheme uses an adaptive fine-tuning algorithm. When the standard deviation of the interference does not exceed the threshold, the cross entropy algorithm is adopted to update the parameters of the deep network architecture; when the standard deviation of the interference exceed the threshold, the mean squared algorithm is adopted to update the parameters of the deep network architecture. As for the on-line phase, the Bayesian probability algorithm is used for position estimation. The experiments were implemented and validated in a noisy and noise-freed indoor environment. The experimental results show that the proposed scheme can improve the accuracy of the training data and reduce the localization error.

1.Abstract
2.Introduction
3.Related Works
　3.1 Related works
　3.2 Motivation
　3.3 Contribution
4.Preliminaries
　4.1 System architecture
　4.2 Problem formulation
　4.3 Basic idea
5.The Adaptive Wi-Fi Indoor Localization Scheme in the Deep Belief Net-
works
　5.1 The Off-line Phase
　5.2 The On-line Phase
6.Simulation Results
　6.1 The data accuracy
　6.2 The localization errors
　6.3 The Training time
7.Conclusions
8.Acknowledgments

[1] P.Smolensky,“Information Processing in Dynamical Systems: Foundations of Harmony Theory,”in MIT Press, pp.194–281,1986.
[2] G.E.Hinton,“Training Products of Experts by Minimizing Contrastive Divergence,”pp. 1771–1800,Aug.2002.
[3] D.ackerly, W.Mendenhall,and R.L.Scheaffer,“Mathematical Statistics with Applications (7ed.),”2008.
[4] P.deBoer,D.Kroese,S.Mannor,and R.Y.Rubinstein,“A Tutorialon the Cross-Entropy Method,”pp.19–67,Jan.2005.
[5] L.Ma,J.Li,Y.Xu,and W.Meng,“Radio Map Recovery and Noise Reduction Method for Green WiFi Indoor Positioning System Based on Inexact Augmented Lagrange Multiplier Algorithm,” in Proceedings of IEEE International Conference on Global Communications (GLOBECOM 2015),pp.1–5,San Diego,CA,USA,Dec.2015.
[6] M.Pajovic,P.Orlik,T.Koike-Akino,K.J.Kim,H.Aikawa,andT.Hori,“An Unsupervised Indoor Localization Method Based on Received Signal Strength (RSS) Measurements,”in Proceedings of IEEE International Conference on Global Communications (GLOBECOM 2015),pp.1–6,San Diego,CA,USA,Dec.2015.
[7] D.Li,B.Zhang,K.Huang,and C.Li,“Measurement-Based Access Point Deployment Mechanism for Indoor Localization,”in Proceedings of IEEE International Conference on Global Communications (GLOBECOM 2015),pp.1–6,San Diego,CA,USA,Dec.2015.
[8] B.Su,J.Cui,A.Li,and Y.Chang,“Performance Study on the ECID Positioning Enhancement in LTE-A Het-Net with RRH,”in Proceedings of IEEE International Conference on 82nd Vehicular Technology (VTC Fall2015),pp.1–5,Boston,USA,Sept.2015.
[9] A.Alonazi,Y.Ma,and R.Tafazolli,“Less-Calibration Wi-Fi based Indoor Positioning,”in Proceedings of International Conference on Communications (ICC 2015),pp.2733–2738, London,UK,June.2015.
[10] X.Wang,L.Gao,and S.Mao,“PhaseFi: Phase Fingerprinting for Indoor Localization with A Deep Learning Approach,”in Proceedings of IEEE International Conference on Global Communications (GLOBECOM 2015),pp.1–6,San Diego,CA,USA,Dec.2015.
[11] X.Wang,L.Gao,S.Mao,and S.Pandey,“DeepFi: Deep Learning for Indoor Fingerprinting Using Channel State Information,”in Proceedings of IEEE International Conference on Wireless Communications and Networking (WCNC2015),pp.1666–1671,New Orleans, LA USA,June.2015.
[12] Y.Gu,Y.Chen,J.Liu,and X.Jiang,“Online Deep Intelligence for Wi-Fi
Indoor Localization,”in in ACM International Joint Conference on Pervasive and ACM International Symposium on Wearable Computers (ACM2015),pp.2932,Sept.2015.
[13] G.Félix,M.Siller,and E.N.Álvarez,“A Fingerprinting Indoor Localization Algorithm Based Deep Learning,”in Proceedings of IEEE International Conference on Ubiquitous and Future Networks (ICUFN 2016),p.1006–011,Austria,August.2016.
[14] W.Zhang,K.Liu,W.Zhang,Y.Zhang,and J.Gu,“Wi-Fi Positioning Based On Deep Learning,” in Proceedings of IEEE International Conference on Information and Automation (ICIA 2014),p.1176–1179,Hailar,China,July.2014.
[15] W.Zhang,K.Liu,W.Zhang,Y.Zhang,and J.Gu,“Deep Neural Networks for Wireless
Localization in Indoor and Outdoor Environments,”pp.279–287,June.2016.
[16] Y.Wen,X.Tian,X.Wang,and S.Lu,“Fundamental Limits of RSS Fingerprinting Based Indoor Localization,”in Proceedings of IEEE International Conference on Computer Communications (INFOCOM2015),pp.2479–2487,Hong Kong,May.2015.
[17] S. He,S.H,G.Chan,L.Yu,and N.Liu,“Fusing Noisy Fingerprints with Distance
Bounds for Indoor Localization,”in Proceedings of IEEE International Conference on Computer Communications (INFOCOM2015),pp.2506–514,HongKong,May.2015.
[18] M.Zhou,Q.Pu,K.Xu,X.Huang,and Z.Tian,“Location Fingerprint Discrimination
Maximization for Indoor WLAN Access Point Optimization Using Fast Discrete Water Filling,” in Proceedings of IEEE International Conferenceon Global Communications (GLOBECOM 2015),pp.1–6,San Diego,CA,USA,Dec.2015.
[19] J.Niu,B.Wang,L.Cheng,and J.J.P.C.Rodrigues,“WicLoc: An Indoor Localization System Based on WiFi Fingerprints and Crowd sourcing,”in Proceedings of IEEE International Conference on Communications (ICC 2015),pp.3008–3013,London,UK,June.2015.
[20] S.Shi,S.Sigg,and Y.Ji,“Probabilistic Fingerprinting Based Passive Device Free Localization from Channel State Information,”in Proceedings of IEEE International Conference on 83rd Vehicular Technology (VTC Spring2015),pp.1-5,Nanjing,China,May.2015.