研究生: |
鄭和軒 Han-Hsuan Cheng |
---|---|
論文名稱: |
融合WiFi訊號強度與人體姿態估計進行兩階段定位系統 Two-Phase Positioning System Based on the Fusion of Wi-Fi Signal Strength and Pose Estimation |
指導教授: |
呂政修
Jenq-Shiou Leu |
口試委員: |
陳維美
Wei-Mei Chen 阮聖彰 Shanq-Jang Ruan 周承復 Cheng-fu Chou 呂政修 Jenq-Shiou Leu |
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: | 2022 |
畢業學年度: | 110 |
語文別: | 中文 |
論文頁數: | 52 |
中文關鍵詞: | 機器學習 、姿態估計 、位置感知 、室內導航 、WiFi 位置估計 |
外文關鍵詞: | Machine learning, Pose estimation, Location awareness, Indoor navigation, WiFi positioning |
相關次數: | 點閱:281 下載:0 |
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 |
由於近年來人們對於定位的重視,全球定位系統(Global Positioning System, GPS)已被廣泛使用於我們生活中的應用,卻礙於建築物的干擾訊號傳播導致GPS在室內定位並不準確,因此如何在室內達到高精度定位成為人們重視的研究議題,傳統的方法是以訊號強度為基礎如:藍牙、Wi-Fi、ZigBee,通過三邊測量估算裝置位置,然而,基於訊號的定位方法容易因為室內環境的多路徑干擾,導致環境中的訊號分佈變動性大,產生高定位誤差,而近年來深度學習的蓬勃發展使研究人員藉由成熟的影像辨識技術對行人進行位置估計與室內定位,卻無法得到設備資訊以識別人員身份,為此我們提出了一種基於Wi-Fi與影像的高精度人員室內定位方法。
室內定位系統分為兩階段定位,第一階段通過使用智慧型手機收集三台Wi-Fi基地台兩個頻段2.4GHz及5GHz的訊號接受強度,並以機器學習方法進行粗精度定位預測,接著在第二階段分析監視攝影機捕捉的人員畫面,並以姿態估計模型提取影像中行人們的腳點座標,再藉由直接線性轉換與線性回歸模型得到影像人員的位置,最後與第一階段的Wi-Fi定位位置進行匹配,完成可識別人員的室內定位系統。
本研究採用的實驗場域具備多遮蔽物及訊號干擾,因此我們收集2.4GHz及5GHz兩個頻段的訊號接受強度,減少2.4GHz的訊號干擾以實現更高的Wi-Fi定位精度,Wi-Fi的平均定位誤差達1.4公尺,並分析兩個頻段的定位表現。在影像定位方面我們則提出兩種用於影像中的行人腳點提取方法,並以機器學習模型減少因為鏡頭扭曲與直接線性轉換造成的誤差,結果表明我們改善後的腳點提取方法能夠降低50%的定位誤差,也指出通過機器學習模型預測的定位結果比僅以2D線性變換的誤差減少約0.4公尺,達到誤差0.4公尺的高精度室內定位。
Due to the importance of positioning in recent years, Global Positioning System (GPS) has been widely used in our life applications, whereas due to the interference of signal propagation in buildings, GPS is not accurate in indoor positioning. Therefore, how to achieve high precision positioning indoors has become an important research issue. However, the signal-based positioning method is prone to high positioning error due to the multipath interference in the indoor environment, which leads to high variability of the received signal. To this end, we propose a Wi-Fi and image-based method for high-precision indoor positioning for clients.
The first phase of the system is scheduled for using smartphones to collect the received signal strength of three Wi-Fi base stations in two frequency bands, 2.4GHz and 5GHz, and use machine learning methods to perform coarse accuracy location prediction. In the second stage we analyze the images of people captured by surveillance cameras and extract the foot position of pedestrians through pose estimation model. In order to transform the position of the feet from the camera perspective to the floor plane, we use direct linear transformation and linear regression model to obtain the more precise coordinate. The location of the person in the image is then matched with the Wi-Fi location from the smartphone to complete an indoor location system that could identify the person.
In this study, we collected the received signal strengths of 2.4GHz and 5GHz bands to reduce the signal interference, and achieved higher Wi-Fi positioning accuracy with an average distance error 1.4 meters, and analyzed the positioning performance of two bands. In terms of pedestrian estimation, we propose two methods to extract pedestrian foot position in image and use machine learning models to reduce the errors caused by lens distortion and direct linear transformation. Our result also shows that the predicted positioning result by the machine learning model is 0.4m less than the error of 2D direct linear transform only, achieving 0.4 m high accuracy positioning.
[1] C. Rose, J. Britt, J. Allen and D. Bevly, "An Integrated Vehicle Navigation System Utilizing Lane-Detection and Lateral Position Estimation Systems in Difficult Environments for GPS," in IEEE Transactions on Intelligent Transportation Systems, vol. 15, no. 6, pp. 2615-2629, Dec. 2014
[2] ““Ericsson Mobility Report 2020” available at: https://www.ericsson.com/assets/local/reports-papers/mobility-report/documents/2020/june2020-ericsson-mobility-report.pdf
[3] A. A. Sohan, M. Ali, F. Fairooz, A. I. Rahman, A. Chakrabarty and M. R. Kabir, "Indoor Positioning Techniques using RSSI from Wireless Devices," 2019 22nd International Conference on Computer and Information Technology (ICCIT), 2019, pp. 1-6, doi: 10.1109/ICCIT48885.2019.9038591.
[4] L. Pei, R. Chen, J. Liu, T. Tenhunen, H. Kuusniemi and Y. Chen, "An Inquiry-based Bluetooth indoor positioning approach for the Finnish pavilion at Shanghai World Expo 2010," IEEE/ION Position, Location and Navigation Symposium, 2010, pp. 1002-1009, doi: 10.1109/PLANS.2010.5507274.
[5] G. Schroeer, "A Real-Time UWB Multi-Channel Indoor Positioning System for Industrial Scenarios," 2018 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2018, pp. 1-5, doi: 10.1109/IPIN.2018.8533792.
[6] T. Wu, L. -K. Chen and Y. Hong, "A vision-based indoor positioning method with high accuracy and efficiency based on self-optimized-ordered visual vocabulary," 2016 IEEE/ION Position, Location and Navigation Symposium (PLANS), 2016, pp. 48-56, doi: 10.1109/PLANS.2016.7479682.
[7] T. Tsai, C. Chang and S. Chen, "Vision based indoor positioning for intelligent buildings," 2016 2nd International Conference on Intelligent Green Building and Smart Grid (IGBSG), 2016, pp. 1-4, doi: 10.1109/IGBSG.2016.7539419.
[8] “The Growth in Connected IoT Devices is Expected to Generate 79.4ZB of Data in 2025, According to a New IDC Forecast” available at: https://www.businesswire.com/news/home/20190618005012/en/The-Growth-in-Connected-IoT-Devices-is-Expected-to-Generate-79.4ZB-of-Data-in-2025-According-to-a-New-IDC-Forecast
[9] Y. Wang, Xu Yang, Yutian Zhao, Yue Liu and L. Cuthbert, "Bluetooth positioning using RSSI and triangulation methods," 2013 IEEE 10th Consumer Communications and Networking Conference (CCNC), 2013, pp. 837-842, doi: 10.1109/CCNC.2013.6488558.
[10] Jooyoung Kim, Myungin Ji, Ju-il Jeon, Sangjoon Park and Y. Cho, "K-NN based positioning performance estimation for fingerprinting localization," 2016 Eighth International Conference on Ubiquitous and Future Networks (ICUFN), 2016, pp. 468-470, doi: 10.1109/ICUFN.2016.7537073.
[11] J. Kietlinski - Zaleski, T. Yamazato and M. Katayama, "Experimental validation of TOA UWB positioning with two receivers using known indoor features," IEEE/ION Position, Location and Navigation Symposium, 2010, pp. 505-509, doi: 10.1109/PLANS.2010.5507253.
[12] R. Hach and A. Rommel, "Wireless synchronization in time difference of arrival based real time locating systems," 2012 9th Workshop on Positioning, Navigation and Communication, 2012, pp. 193-195, doi: 10.1109/WPNC.2012.6268763.
[13] Y. Wang and K. C. Ho, "Unified Near-Field and Far-Field Localization for AOA and Hybrid AOA-TDOA Positionings," in IEEE Transactions on Wireless Communications, vol. 17, no. 2, pp. 1242-1254, Feb. 2018, doi: 10.1109/TWC.2017.2777457.
[14] R. Girshick, J. Donahue, T. Darrell and J. Malik, "Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation," 2014 IEEE Conference on Computer Vision and Pattern Recognition, 2014, pp. 580-587, doi: 10.1109/CVPR.2014.81.
[15] Alexey Bochkovskiy, Chien-Yao Wang and Hong-Yuan Mark Liao, “YOLOv4: Optimal Speed and Accuracy of Object Detection” arXiv:2004.10934 [cs],Apr. 2020.
[16] Z. Cao, G. Hidalgo, T. Simon, S. Wei and Y. Sheikh, "OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields" in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 43, no. 01, pp. 172-186, 2021.doi: 10.1109/TPAMI.2019.2929257
[17] KaewTraKulPong P., Bowden R. (2002) An Improved Adaptive Background Mixture Model for Real-time Tracking with Shadow Detection. In: Remagnino P., Jones G.A., Paragios N., Regazzoni C.S. (eds) Video-Based Surveillance Systems. Springer, Boston, MA. https://doi.org/10.1007/978-1-4615-0913-4_11
[18] Chunluan Zhou and Junsong Yuan. Bi-box regression for pedestrian detection and occlusion estimation. In The European Conference on Computer Vision (ECCV), September 2018.
[19] S. Zhang, J. Guo, W. Wang and J. Hu, "Indoor 2.5D Positioning of WiFi Based on SVM," 2018 Ubiquitous Positioning, Indoor Navigation and Location-Based Services (UPINLBS), 2018, pp. 1-7, doi: 10.1109/UPINLBS.2018.8559903.
[20] Lin TY. et al. (2014) Microsoft COCO: Common Objects in Context. In: Fleet D., Pajdla T., Schiele B., Tuytelaars T. (eds) Computer Vision – ECCV 2014. ECCV 2014. Lecture Notes in Computer Science, vol 8693. Springer, Cham. https://doi.org/10.1007/978-3-319-10602-1_48
[21] Karara H. M. Abdel-Aziz, Y. I. Direct linear transformation into object space coordinates in close-range photogrammetry. 1971.
[22] N. A. M. Maung and W. Zaw, "Comparative Study of RSS-based Indoor Positioning Techniques on Two Different Wi-Fi Frequency Bands," 2020 17th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), 2020, pp. 185-188, doi: 10.1109/ECTI-CON49241.2020.9158211.