簡易檢索 / 詳目顯示

研究生: 許任捷
JEN-CHIEH HSU
論文名稱: 利用綜合機器學習演算法實現混合式室內定位導航方案之設計與實作
Design and Implementation of Hybrid Indoor Localization and Navigation Schemes with Ensemble Learning Algorithm
指導教授: 呂政修
Jenq-Shiou Leu
口試委員: 呂政修
Jenq-Shiou Leu
周承復
Cheng-Fu Chou
阮聖彰
Shanq-Jang Ruan
鄭瑞光
Ray-Guang Cheng
林敬舜
Ching-Shun Lin
學位類別: 碩士
Master
系所名稱: 電資學院 - 電子工程系
Department of Electronic and Computer Engineering
論文出版年: 2017
畢業學年度: 105
語文別: 中文
論文頁數: 69
中文關鍵詞: 室內定位導航行人航位推算訊號特徵指紋辨識綜合機器學習混合系統
外文關鍵詞: indoor localization and navigation, pedestrian dead reckoning, fingerprinting, ensemble learning, hybrid systems
相關次數: 點閱:323下載:2
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報
  • 在現今的生活中,隨著智慧型手機和行動網路的普及化,人們開始追求將手機上快捷便利的服務模式應用到其他領域,也因此依賴由定位與導航所衍伸的行動定位服務(Location-Based Service,LBS)需求逐漸提升。然而最普及的全球定位系統(Global Positioning System,GPS)的訊號會被建築物的許多因素所阻礙,進而發展出室內定位的需求。因此智慧型手機透過許多的應用程序利用不同的技術以及感測輸入來進行室內定位。大多數的室內定位(Indoor Positioning)系統都依據從室內無線發射裝置所接收到的接收訊號強度(Received Signal Strength Indicator,RSSI)來當作定位的參考。但是在現實情況下,室內定位系統的精準度容易因訊號干擾而受到影響,所以藉由整合室內定位的各式方案,以促進室內定位的精準,使其更容易得到對應的室內位置資訊。
      在此論文中,我們提出並比較了使用行人航位推算(Pedestrian Dead Reckoning,PDR)技術,接收訊號強度指紋(RSSI Fingerprint),以及綜合機器學習演算法(Ensemble Machine Learning),以及組合式系統,以便可以找到可以輕鬆便宜應用在現今商品現貨設備上的方案,而以上方案的室內定位方式是藉由訊號紋路特徵比對,以及智慧型手機(Smartphone)的感測器輔助,讓其使用者可以得到所在的室內位置。我們所設計的方案在實驗結果中,顯示此些方案是非常有希望改善室內定位的精準度,縮小其定位的誤差距離,並且容易在現有的裝置上實現。


    Nowadays, there is an increasing demand for indoor localization and navigation services with readily available Commercial off-the-shelf (COTS) devices, such as smartphones or wearable devices. Many applications on smartphones exploit different techniques and input for positioning. Most of these systems rely on Received Signal Strengths (RSSs) from indoor wireless emitting devices. However, the accuracy of indoor position systems is easily affected by signal interference in realistic situations.
    In this paper, we propose and compare several indoor localization and navigation systems using Pedestrian Dead Reckoning (PDR), fingerprinting, ensemble machine learning, and hybrid systems, in order to find schemes that can be easily and cheaply applied to COTS devices.

    論文摘要 I ABSTRACT II 誌謝 III 目錄 IV 第1章 緒論 1 1.1 研究背景與動機 1 1.3 章節提要 4 第2章 定位相關技術 5 2.1 定位原理探討 5 2.1.1 收訊時間測量法(Time of Arrival,TOA) 6 2.1.2 收訊時間差測量法(Time Difference of Arrival, TDOA) 6 2.1.3 接收訊號角度測量法(Angle of Arrival,AOA) 7 2.1.4 接收訊號強度(Received Signal Strength Indicator,RSSI) 8 2.1.5 訊號傳輸定位方法的比較 9 2.1.6 三角技術測量法(Triangulation Technique) 10 2.2 定位感測技術 11 2.2.1 全球衛星定位系統(Global Positioning System,GPS) 12 2.2.2 Wi-Fi 定位技術 13 2.2.3 藍牙定位系統(Bluetooth) 14 2.2.4 定位感測技術的比較 15 2.3 無線電波傳遞特性 16 2.3.1 反射與折射(Reflection and Refraction) 16 2.3.2 繞射(Diffraction) 17 2.3.3 散射(Scattering) 17 2.3.4 多重路徑效應(Multipath Effect) 18 第3章 室內定位系統的設計 19 3.1 設計步驟 19 3.2 系統流程 21 3.3系統環境建立 24 3.3.1 佈建iBeacon藍牙訊號 24 3.4 建構接收訊號強度指紋資料庫 25 3.4.1 iBeacon 25 3.4.2 訊號採集 27 3.4.3 訊號濾波處理 27 3.5 定位模型建立 28 3.5.1 行人航位推算(Pedestrian Dead Reckoning) 28 3.5.2 訊號特徵指紋辨識(Fingerprinting) 30 3.5.2.1 K-最近鄰居法(KNN,K-Nearest Neighbor) 31 3.5.2.2 隨機森林演算法(RF,Random Forests) 33 3.5.2.3 類神經網絡(Artificial Neural Network,ANN) 35 3.5.3 綜合機器學習(Ensemble Machine Learning) 37 3.5.4 混合式方案(Hybrid Scheme) 39 第4章 實驗測試與評估結果 41 4.1 室內定位系統裝置介紹 41 4.1.1 智慧型行動裝置 41 4.1.2 iBeacon藍牙發射器 42 4.3 實驗測試環境場地 43 4.3.1 電資學院EE705-6實驗場域之環境建置 44 4.3.2 電資學院EE1F實驗場域之環境建置 44 4.3.3 蒐集訊號 45 4.3.4 定位系統手機應用程式 47 4.4 實驗測試 48 4.4.1 Machine Learning-KNN與RF數值設定 48 4.4.2 執行濾波時間與KNN離線準確率測試 50 4.4.3 Ensemble Machine Learning 離線測試 51 4.4.4 Ensemble Machine Learning實際系統線上測試 52 4.4.5 組合式方案以及其他方案的實驗結果與分析 54 4.4.6 KNN與Ensemble Machine Learning執行環境與耗電量比較 55 第5章 結論 57 參考文獻 58

    [1] M. Altini, D. Brunelli, E. Farella, and L. Benini. Bluetooth indoor localization with multiple neural networks. In Wireless Pervasive Computing (ISWPC), 2010 5th IEEE International Symposium on, pages 295{300. IEEE, 2010.
    [2] J.-S. Leu, M.-C. Yu, and H.-J. Tzeng. Improving indoor positioning precision by using received signal strength _ngerprint and footprint based on weighted ambient wi-_ signals. Computer Networks, 91:329{340, 2015.
    [3] J. Zhu, K. Zeng, K.-H. Kim, and P. Mohapatra. Improving crowd-sourced wi-_ localization systems using bluetooth beacons. In Sensor, Mesh and Ad Hoc Communications and Networks (SECON), 2012 9th Annual IEEE Communications Society Conference on, pages 290{298. IEEE, 2012.
    [4] D. Taniuchi and T. Maekawa, "Robust Wi-Fi based indoor positioning with ensemble learning," 2014 IEEE 10th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob), pp. 592-597, Larnaca, 2014.
    [5] J. S. Leu, M. C. Yu, and H. J. Tzeng, “Improving indoor position precision by using received signal strength fingerprint and footprint based on weighted ambient wi-fi signals,” Computer Networks, vol. 91, pp. 329-340, 2015.
    [6] S. He and S. H. G. Chan, "Wi-Fi Fingerprint-Based Indoor Positioning: Recent Advances and Comparisons," in IEEE Communications Surveys & Tutorials, vol. 18, no. 1, pp. 466-490, 2016.
    [7] Z. Chen, H. Zou, H. Jiang, Q. Zhu, Y. C. Soh, and L. Xie, “Fusion of WiFi, smartphone sensors and landmarks using the Kalman filter for indoor localization,” Sensors, vol. 15, no. 1, pp. 715-732, 2015
    [8] Ge Shi, Yong Li, Li Xi and Baitao Zang, "A robust method for indoor localization based on visible light communication," 2016 2nd IEEE International Conference on Computer and Communications (ICCC), pp. 2154-2158, Chengdu, 2016.
    [9] NIST, “Rules of the PerfLoc Prize Competition”, Mar. 22 to Sep. 21, 2017. [Online]. Avialable: https://perfloc.nist.gov/perfloc-competition-rule.php [Accessed 02- Jun- 2017]
    [10] Microsoft Inc., “2017 Microsoft Indoor Localization Competition @ IPSN 2017 – Official Results”, Apr. 18-20, 2017. [Online]. Aviable: https://www.microsoft.com/en-us/research/event/microsoft-indoor-localization-competition-ipsn-2017/ [Accessed 02- Jun- 2017]

    QR CODE