隨著物聯網(IoT)技術的發展，室內定位服務的需求也逐步提升。其中，基於通道狀態信息(Channel State Information, CSI)進行室內指紋定位的方法成為了熱門的研究領域。然而，通道狀態信息本身有著高雜訊、高複雜度的問題，所以目前還不存在有效的定位方法。同時，許多的指紋定位方法在離線模型訓練上相當耗費時間。為了解決以上問題，我們提出一個基於寬度學習(Broad Learning System , BLS)的室內通道狀態信息指紋定位系統，並透過一連串的方法前處理通道狀態信息，使其更好的被寬度學習模型學習。首先，我們使用了資料塊(Data Nugget)演算法來過濾離群值並產生通道狀態信息的代表，而後使用張量分解(Tensor Decomposition)重構資料。此外，為了減少寬度學習模型的複雜度，我們透過等距特徵映射(Isometric Mapping, Isomap) 提取通道狀態信息的特徵。最終，前處理過的通道狀態信息即是寬度學習模型的輸入。經過實驗證明，我們提出的定位方法在室內環境中，比已有的方法更佳。

With the development of Internet of Things (IoT), the demand for location-based services in indoor environments proliferates. Channel State Information (CSI) based fingerprint localization has become a research hot spot. However, a compelling method to address this issue does not appear because of the hardship of processing CSI data which contain considerable noise and high complexity. Meanwhile, many fingerprint localization algorithms have a time-consuming offline training phase. Therefore, we propose an indoor CSI fingerprint localization system based on the broad learning system (BLS), which conducts a uniform method for CSI data preprocessing to make it better learned by BLS. First, we filter the outliers, and generate delegates of CSI by the data nugget algorithm, then, we use tensor decomposition to reconstruct the CSI delegates. Moreover, we utilize Isometric mapping (Isomap) to extract CSI features to reduce BLS's complexity. Consequently, the preprocessed CSI data is the input for BLS. The experimental results show that our scheme outperforms several existing algorithms in indoor environments.

Abstract ii
1 Introduction 1
1.1 Indoor Localization System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Related Work 4
2.1 Machine learning in Indoor CSI Fingerprint Localization . . . . . . . . . . . . 4
2.2 Dimension Reduction in Indoor CSI Fingerprint Localization . . . . . . . . . 6
2.3 Indoor Fingerprint Localization Using UAV . . . . . . . . . . . . . . . . . . . 6
3 System Model 7
3.1 Channel State Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Feature Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.3 Architecture of System Model . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4 Indoor CSI fingerprint Localization (ICLoc) 10
4.1 Data Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.2 BLS Training Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.3 Localization Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5 Experiments 18
5.1 Experiment Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.2 Impact of the number of CSI delegates . . . . . . . . . . . . . . . . . . . . . . 20
5.3 Impact of the number of features . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.4 Classification Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.4.1 Testing the RPs from the Training Set . . . . . . . . . . . . . . . . . . 22
5.4.2 Testing the RPs from the Testing Set . . . . . . . . . . . . . . . . . . . 23
5.5 Localization Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.5.1 Testing the RPs inside the Training Area . . . . . . . . . . . . . . . . . 24
5.5.2 Testing the RPs outside the Training Area . . . . . . . . . . . . . . . . 24
5.5.3 Testing the RPs of Trajectory . . . . . . . . . . . . . . . . . . . . . . 26
5.5.4 Performance Evaluation with the Help of UAV . . . . . . . . . . . . . 27
6 Conclusion 29
References 30
iii

1] K. Jo, K. Chu, and M. Sunwoo, “Interacting multiple model filter-based sensor fusion
of GPS with in-vehicle sensors for real-time vehicle positioning,” IEEE Transactions on
Intelligent Transportation Systems (T-ITS), vol. 13, no. 1, pp. 329–343, 2012.
[2] C. Gao, G. Zhao, and H. Fourati, “Cooperative localization and navigation: Theory, Re-
search, and Practice,” USA: CRC Press, 2019.
[3] D. Mototolea, I. Nicolaescu, A. Mîndroiu, and A. Vlăsceanu, “Evaluation of errors caused
by inaccurate clock synchronization in time difference of arrival-based localization sys-
tems,” International Symposium on Signals, Circuits and Systems (ISSCS), Iasi, Romania,
2019, pp. 1-4.
[4] C.-M. Own, J. Hou, and W. Tao, “Signal fuse learning method with dual bands wifi signal
measurements in indoor positioning,” IEEE Access, vol. 7, pp. 131805 – 131817, 2019.
[5] H.-C. Tsai, C.-J. Chiu, P.-H. Tseng, and K.-T. Feng, “Refined autoencoder-based CSI hid-
den feature extraction for indoor spot localization,” IEEE 88th Vehicular Technology Con-
ference (VTC-Fall), Chicago, IL, USA 2018, pp. 1-5.
[6] D. Liu, Z. Liu, and Z. Song, “LDA-based CSI amplitude fingerprinting for device-free
localization,” Chinese Control And Decision Conference (CCDC), Hefei, China, 2020,
pp. 2020-2023.
[7] Z. Gao, Y. Gao, S. Wang, D. Li, and Y. Xu, “CRISLoc: Reconstructable CSI fingerprint-
ing for indoor smartphone localization,” IEEE Internet of Things Journal, vol. 8, no. 5,
pp. 3422–3437, 2021.
[8] J. Wang and J. G. Park, “An enhanced indoor ranging method using CSI measurements
with extended Kalman filter,” IEEE/ION Position, Location and Navigation Symposium
(PLANS), Portland, Oregon, 2020, pp. 1342-1348.
[9] W. Xun, L. Sun, C. Han, Z. Lin, and J. Guo, “Depthwise separable convolution based
passive indoor localization using CSI fingerprint,” IEEE Wireless Communications and
Networking Conference (WCNC), Seoul, Korea, 2020, pp. 1-6.
[10] X. Zhu, T. Qiu, W. Qu, X. Zhou, M. Atiquzzaman, and D. Wu, “BLS-Location: A wireless
fingerprint localization algorithm based on broad learning,” IEEE Transactions on Mobile
Computing, pp. 1–1, 2021.
[11] Z. Liu, D. Liu, J. Xiong, and X. Yuan, “A parallel adaboost method for device-free indoor
localization,” IEEE Sensors Journal, vol. 22, no. 3, pp. 2409–2418, 2022.
[12] Y. Long, L. Xie, M. Zhou, and Y. Wang, “Indoor CSI fingerprint localization based on
tensor decomposition,” IEEE/CIC International Conference on Communications in China
(ICCC), Portland, Oregon, USA, 2020, pp. 1190-1195.
[13] F. Gringoli, M. Schulz, J. Link, and M. Hollick, “Free your CSI: A channel state infor-
mation extraction platform for modern Wi-Fi chipsets,” 13th International Workshop on
Wireless Network Testbeds, Experimental Evaluation & Characterization, 2019.
[14] B.-Y. Chang and J.-P. Sheu, “Indoor localization with CSI fingerprint utilizing depthwise
separable convolution neural network,” IEEE International Symposium on Personal, In-
door and Mobile Radio Communications (PIMRC), Virtual Conference, 2022, pp. 1-6.
[15] A. Foliadis, M. H. C. Garcia, R. A. Stirling-Gallacher, and R. S. Thomä, “CSI-based lo-
calization with CNNs exploiting phase information,” IEEE Wireless Communications and
Networking Conference (WCNC), Nanjing, China, 2021, pp. 1-6.
[16] S. S. Moosavi and P. Fortier, “A fingerprint localization method in collocated massive
MIMO-OFDM systems using clustering and gaussian process regression,” IEEE 92nd
Vehicular Technology Conference (VTC2020-Fall), Victoria, BC, Canada, 2020, pp. 1-5.
[17] B. O. Ayinde, T. Inanc, and J. M. Zurada, “Regularizing deep neural networks by enhanc-
ing diversity in feature extraction,” IEEE Transactions on Neural Networks and Learning
Systems, vol. 30, no. 9, pp. 2650–2661, 2019.
[18] N. Kardan and K. O. Stanley, “Mitigating fooling with competitive overcomplete out-
put layer neural networks,” International Joint Conference on Neural Networks (IJCNN),
Anchorage, AK, USA, 2017, pp. 518-525.
[19] K. P. Singh, R. Bhai, V. Mishra, P. Nagar, and J. Kasinayal, “Localization in wireless
sensor network using LLE-ISOMAP algorithm,” TENCON 2017 - 2017 IEEE Region 10
Conference, Penang, Malaysia, 2017, pp. 393-397.
[20] A. C. Neto and A. L. M. Levada, “ISOMAP-KL: a parametric approach for unsupervised
metric learning,” 33rd SIBGRAPI Conference on Graphics, Patterns and Images (SIB-
GRAPI), virtual conference, 2020, pp. 287-294.
[21] B. Xiang, F. Yan, Y. Zhu, T. Wu, W. Xia, J. Pang, W. Liu, G. Heng, and L. Shen, “UAV
assisted localization scheme of wsns using RSSI and CSI information,” IEEE 6th Inter-
national Conference on Computer and Communications (ICCC), Chengdu, China, 2020,
pp. 718-722.
[22] C. Wu, T. Qiu, C. Zhang, W. Qu, and D. O. Wu, “Ensemble strategy utilizing a broad learn-
ing system for indoor fingerprint localization,” IEEE Internet of Things Journal, vol. 9,
no. 4, pp. 3011–3022, 2022.
[23] S. M. Hamdi and R. Angryk, “Interpretable feature learning of graphs using tensor de-
composition,” IEEE International Conference on Data Mining (ICDM), Beijing, China,
2019, pp. 270-279.
[24] T. Kolda and B. Bader, “Tensor decompositions and applications,” SIAM Review, vol. 51,
pp. 455–500, 08 2009.
[25] M. A. A. Cox and T. F. Cox, “Multidimensional scaling,” Monographs on Statistics and
Applied Probability, vol. 88, 2001.
[26] C. L. P. Chen and Z. Liu, “Broad learning system: An effective and efficient incremental
learning system without the need for deep architecture,” IEEE Transactions on Neural
Networks and Learning Systems, vol. 29, no. 1, pp. 10–24, 2018.
[27] S. Wu, G. Li, L. Deng, L. Liu, D. Wu, Y. Xie, and L. Shi, “l1 -norm batch normalization
for efficient training of deep neural networks,” IEEE Transactions on Neural Networks
and Learning Systems, vol. 30, no. 7, pp. 2043–2051, 2019.
[28] G. Klambauer, T. Unterthiner, A. Mayr, and S. Hochreiter, “Approximation by superpo-
sitions of a sigmoidal function,” Mathematics of Control, Signals, and Systems (MCSS),,
vol. 2, no. 4, pp. 303–314, 1989.
[29] G. Klambauer, T. Unterthiner, A. Mayr, and S. Hochreiter, “Self-normalizing neural
networks,” Advances in Neural Information Processing Systems 30 (NIPS 2017), Long
Beach, California, USA, 2017, pp. 972–981.
[30] X. Li, Z. Hu, and X. Huang, “Combine relu with tanh,” IEEE 4th Information Technology,
Networking, Electronic and Automation Control Conference (ITNEC), Chongqing, China,
2020, pp. 51-55.