基於Wi-Fi與IMU 的精準室內定位系統

An Accurate Indoor Positioning System Based On Wi-Fi And IMU

王韻豪(Yun-Hao Wang)

指導教授 ： 周承復

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2021年)

https://doi.org/10.6342/NTU202103324

摘要

近年來，物聯網設備的數量呈指數級增加，人們對於Wi-Fi的需求也隨之增長，這間接導致基於Wi-Fi指紋的定位服務快速發展。然而，由於Wi-Fi訊號的不穩定性、設備的異質性以及環境的動態變化，這些因素大大降低了傳統Wi-Fi指紋定位演算法的精確度與魯棒性。為緩和這些議題，我們提出了基於去噪監督式自動編碼器(Denoising Supervised Autoencoder) 的Wi-Fi指紋定位系統---Wi-Fi DSAR，Wi-Fi DSAR能有效地減少這些因素的影響；再者，我們還提出了新穎的混合式定位系統，它結合了慣性測量單元(Inertial Measurement Unit, IMU)定位與Wi-Fi指紋定位，其主要想法為(1)在Wi-Fi覆蓋薄弱的地方透過慣性測量單元來輔助定位和(2)透過Wi-Fi指紋定位來消除慣性測量單元定位的累積誤差。
實驗結果表明Wi-Fi DSAR不僅對Wi-Fi雜訊與設備異質性具有高度抗性，還能避免過度擬合指紋資料庫，並且達到令人滿意的精確度和魯棒性；再者，我們提出的混合式定位系統，證實有效互補兩種定位方式的缺點，進而達到更高的定位精確度。此外，本篇論文也提供了完整的、精確的資料集，其中包含了Wi-Fi指紋以及慣性測量單元的感測資料，該資料集於國立台灣大學資訊工程學系五樓收集而成，我們將此資料集公開，幫助混合式定位的發展。

關鍵字

室內定位 ； WiFi 指紋定位 ； 慣性測量單元導航 ； 去噪監督式自動編碼器 ； 資料融合

並列摘要

In recent years, the number of IoT devices has grown exponentially, and the demand for Wi-Fi has increased accordingly, which led to the rapid development of indoor positioning services based on Wi-Fi fingerprints. Due to the instability of Wi-Fi received signal strength indication (RSSI), device heterogeneity and dynamic changes in the environment, the accuracy and robustness of traditional Wi-Fi fingerprint-based positioning algorithms usually degrade. To alleviate these issues, we propose a Wi-Fi fingerprint-based positioning system using Denoising Supervised Autoencoder we named Wi-Fi DSAR. Wi-Fi DSAR can effectively reduce the influence of these factors; In addition, we also propose a novel hybrid positioning system, which combines Inertial Measurement Unit (IMU) positioning and Wi-Fi fingerprint-based positioning. The main idea is (1) using IMU to assist positioning in places where Wi-Fi coverage is poor and (2) eliminating the accumulation of errors caused by IMU positioning through Wi-Fi fingerprint-based positioning.
The experimental results show that Wi-Fi DSAR is not only highly resistant to the noise in received Wi-Fi signal and device heterogeneity, but also avoids overfitting the fingerprint database, and achieves satisfactory accuracy and robustness; furthermore, the hybrid positioning system we proposed proves to effectively complement the shortcomings of the two positioning methods, thereby achieving higher positioning accuracy. In addition, this paper also provides a complete and accurate dataset, including Wi-Fi fingerprints and IMU sensing data. The dataset is collected on the fifth floor of the Department of Computer Science and Information Engineering, National Taiwan University. We make this dataset public to help the development of hybrid positioning.

並列關鍵字

Indoor Localization ； WiFi Fingerprinting ； IMU Navigation ； Denoising Supervised Autoencoder ； Data Fusion

參考文獻

M. Abbas, M. Elhamshary, H. Rizk, M. Torki, and M. Youssef. Wideep: Wifibased accurate and robust indoor localization system using deep learning. In 2019 IEEE International Conference on Pervasive Computing and Communications (PerCom, pages 1–10, 2019.

P. Bahl and V. Padmanabhan. Radar: an inbuilding rfbased user location and tracking system. In Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064), volume 2, pages 775–784 vol.2, 2000.

Q. Chang, S. Van de Velde, W. Wang, Q. Li, H. Hou, and S. Heidi. Wifi fingerprint positioning updated by pedestrian dead reckoning for mobile phone indoor localization. In J. Sun, J. Liu, S. Fan, and X. Lu, editors, China Satellite Navigation Conference (CSNC) 2015 Proceedings: Volume III, pages 729–739, Berlin, Heidelberg, 2015. Springer Berlin Heidelberg.

C. Chen, X. Lu, A. Markham, and N. Trigoni. Ionet: Learning to cure the curse of drift in inertial odometry, 2018.

C. Chen, P. Zhao, C. X. Lu, W. Wang, A. Markham, and N. Trigoni. Deep learning based pedestrian inertial navigation: Methods, dataset and ondevice inference, 2020.