臺灣博碩士論文加值系統

本研究基於物聯網架構，設計一移動式空氣汙染感測單元進行都會區細懸浮微粒的濃度監測。此單元由NodeMCU-32S微控制器進行開發，配置PMS5003-G5(懸浮微粒感測模組)及Ublox NEO-6M V2(GPS定位模組)。感測單元透過3G與4G等電信網路架構，將汙染濃度數據及汙染定位座標等資訊回傳至後端伺服器端進行資料蒐集。此移動式感測系統，可改善政府管理的固定式監測站資料相對稀疏的問題，能夠更精確的描述都會生活圈各區域的汙染程度。
其次，本研究運用機器學習模型，結合固定式監測站與本研究提出的移動式感測系統，來估測新莊、三重和蘆洲三大區域懸浮微粒濃度分布情形。本研究採用了四種學習模型進行訓練並評估其結果，分別為決策樹(Decision tree)、隨機森林(Random Forest)、多層感知機(Multilayer Perceptron)、徑向基函數神經網路(Radial Basis Function Neural Network)。以均方根誤差(RMSE)作為指標，決策樹於訓練集的誤差為0.024211，於測試集的誤差為0.029622。隨機森林於訓練集的誤差為0.016763，於測試集的誤差為0.023564。多層感知機於訓練集的誤差為0.089178，於測試集的誤差為0.089928。徑向基函數神經網路於訓練集的誤差為0.083023，於測試集的誤差為0.087194。由學習結果發現隨機森林模型在訓練集以及測試集上表現結果皆為最佳。
為了驗證學習模型的泛化能力，本研究選擇三日進行實地驗證，2019/02/15、2019/02/28及2019/03/01。決策樹模型的誤差值分別為5.877359(μg/m3)、4.332552(μg/m3)及5.824783(μg/m3)。隨機森林模型的誤差值分別為4.671839(μg/m3)、4.561410(μg/m3)及4.578904(μg/m3)。多層感知機模型的誤差值分別為7.594026(μg/m3)、4.353507(μg/m3)及8.350137(μg/m3)。徑向基函數神經網路模型的誤差值分別為7.079262(μg/m3)、3.960256(μg/m3)及15.162552(μg/m3)。結果表明隨機森林模型擁有最穩定且精確的預測值，並能明確標示高汙染區域的分布情形。
本研究預測模型之結果，能夠透過網頁應用程序的方式進行可視化，透過地圖形式讓使用者能更為直觀明白汙染區域分布。

This thesis designs a mobile air pollution sensing system to monitor the concentration of particulate matter 2.5 in the metropolitan area based on the Internet of Things. This system is developed by NodeMCU-32S microcontroller and equipped with PMS5003-G5 (particulate matter sensing module) and Ublox NEO-6M V2 (GPS positioning module). The sensing system transmits data such as concentration of pollution and positioning coordinates to the server for data collection by 3G and 4G. The measurement system can improve the weakness that sparse data of government-managed fixed monitoring stations, and can more accurately describe the pollution levels of various areas of the metropolitan.
This study uses machine learning model, combined with the data from fixed monitoring stations and mobile sensing system to estimate the concentration distribution of particulate matter 2.5 in Xinzhuang, Sanchong and Luzhou. The estimation system uses four regression models which contain Decision Tree (DT), Random Forest (RF), Multilayer Perceptron (MLP) and Radial Basis Function Neural Network (RBFNN). The root mean square error (RMSE) of DT in the training set is 0.024211 and in the testing set is 0.029622. The RMSE of RF in training set is 0.016763 and in the testing set is 0.023564. The RMSE of MLP in training set is 0.089178 and in the testing set is 0.089928. The RMSE of RBFNN in training set is 0.083023 and in the testing set is 0.087194. The results show that RF is better than other regression models.
In order to verify the generalization ability of the learning model, this study selects three days for field verifications, on 2019/02/15, 2019/02/28 and 2019/03/01. The Mean Absolute Error (MAE) of DT are 5.877359, 4.332552 and 5.824783, respectively. The MAE of RF are 4.671839, 4.561410, and 4.578904, respectively. The MAE of MLP are 7.594026, 4.353507, and 8.350137, respectively. The MAE of RBFNN are 7.079262, 3.960256, and 15.162552, respectively. The results show that the RF model has the most stable and accurate predictions and clearly indicates the distribution of highly polluted areas. The results of the predictions can be visualized through the web application, and the map form allows the user intuitively understand the distribution of the contaminated area.

目 錄

摘 要 i
ABSTRACT iii
誌 謝 v
目 錄 vi
表目錄 x
圖目錄 xi
第一章 緒論 1
1.1研究動機 1
1.2研究目的 2
1.3文獻探討 2
1.4大綱 5
第二章 物聯網與細懸浮微粒監測 6
2.1物聯網 6
2.1.1物聯網背景 6
2.1.2物聯網架構 7
2.1.3無線傳輸技術 8
2.2細懸浮微粒 10
2.2.1懸浮微粒定義 10
2.2.2台灣各區域細懸浮微粒來源 12
2.3物聯網應用於細懸浮微粒監測 15
第三章 機器學習 17
3.1前言 17
3.2決策樹演算法 19
3.2.1決策樹演算法介紹 19
3.2.2 CART決策樹運作原理 20
3.3隨機森林演算法 24
3.3.1隨機森林演算法介紹 24
3.3.2隨機森林運作原理 25
3.4多層感知機演算法 26
3.4.1多層感知機介紹 26
3.4.2多層感知機架構 27
3.4.3多層感知機運作原理 30
3.4.4梯度消失現象 34
3.5徑向基函數神經網路 36
3.5.1徑向基函數神經網路介紹 36
3.5.2徑向基函數神經網路運作原理 37
第四章 系統架構與硬體設備 40
4.1系統架構 40
4.1.1環境感測單元 41
4.1.2無線傳輸通道 43
4.1.3雲端資料庫建置 43
4.1.4 Web應用程序 45
4.2硬體設備 46
4.2.1微控制器NodeMCU-32S 46
4.2.2懸浮微粒感測模組 47
4.2.3 GPS定位模組 49
4.2.4硬體設備安裝 49
第五章 細懸浮微粒預測系統設計 51
5.1前言 51
5.2細懸浮微粒污染資料數據集 51
5.3機器學習模型架構 53
5.4模型訓練流程 54
5.4.1資料預處理 55
5.4.2數據分組 56
5.4.3模型訓練 57
5.4.4模型比較 59
5.4.5模型儲存 59
5.5模型結果比較 59
5.6本章結論 65
第六章 實驗結果 66
6.1前言 66
6.2實際預測 66
6.2.1實測資訊 66
6.2.2模型預測結果 68
6.3網頁應用程序 76
6.4本章結論 79
第七章 結論與未來研究方向 80
7.1結論 80
7.2本研究之貢獻 80
7.3未來研究方向 81
參考文獻 82
作者簡介 85

[1]A. Fazziki, D. Benslimane, A. Sadiq, J. Ouarzazi and M. Sadgal, “An Agent Based Traffic Regulation System for The Roadside Air Quality Control,” IEEE Access, vol. 5, 2017, pp. 13192-13201.
[2]K. Hu, A. Rahman, H. Bhrugubanda and V. Sivaraman, “Hazeest: Machine Learning Based Metropolitan Air Pollution Estimation from Fixed and Mobile Sensors,” IEEE Sensors Journal, vol. 17, no. 11, 2017, pp. 3517-3525.
[3]A. Kadri, E. Yaacoub, M. Mushtaha and A. Abu-Dayya, “Wireless Sensor Network for Real-time Air Pollution Monitoring,” IEEE International Conference on Communications, Signal Processing and their Applications, 2013, pp. 1-5.
[4]M. Mead, O. Popoola, G. Stewart, P. Landshoff, M. Calleja, M. Hayes, J. Baldovi, M. McLeod, T. Hodgson, J. Dicks, A. Lewis, J. Cohen, R. Baron, J. Saffell and R. Jones, “The Use of Electrochemical Sensors for Monitoring Urban Air Quality in Low-cost, High-density Networks,” Atmospheric Environment, vol. 70, 2013, pp. 186-203.
[5]B. Predić, Z. Yan, J. Eberle, D. Stojanovic and K. Aberer, “Exposuresense: Integrating Daily Activities with Air Quality Using Mobile Participatory Sensing,” IEEE International Conference on Pervasive Computing Communications Workshops, 2013, pp. 303-305.
[6]K. Shaban, A. Kadri and E. Rezk, “Urban Air Pollution Monitoring System with Forecasting Models,” IEEE Sensors Journal, vol. 16, no. 8, 2016, pp. 2598-2606.
[7]D. Hasenfratz, O. Saukh, C. Walser, C. Hueglin, M. Fierz and L. Thiele, “Pushing The Spatio-temporal Resolution Limit of Urban Air Pollution Maps,” IEEE International Conference on Pervasive Computing and Communications, 2014, pp. 69-77.
[8]M. Bacco, F. Delmastro, E. Ferro and A. Gotta, “Environmental Monitoring for Smart Cities,” IEEE Sensors Journal, vol. 17, no. 23, 2017, pp. 7767-7774.
[9]L. J. Chen, Y. H. Ho, H. H. Hsieh, S. T. Huang, H. C. Lee and S. Mahajan, “ADF: An Anomaly Detection Framework for Large-scale PM2.5 Sensing Systems,” IEEE Internet of Things Journal, vol. 52, no. 2, 2018, pp. 559-570.
[10]V. Singh, C. Carnevale, G. Finzi, E. Pisoni and M. Volta, “A Cokriging Based Approach to Reconstruct Air Pollution Maps Processing Measurement Station Concentrations and Deterministic Model Simulations,” Environmental Modeling and Software, vol. 26, no. 6, 2011, pp. 778-786.
[11]V. Sivaraman, J. Carrapetta, K. Hu and B. G. Luxan, “HazeWatch: A Participatory Sensor System for Monitoring Air Pollution in Sydney,” IEEE Conference on Local Computer Networks Workshops, 2013, pp. 56-64.
[12]H. Qu, W. I. Chan, A. Xu, K. L. Chug, K. H. Lau and P. Guo, “Visual Analysis of The Air Pollution Problem in Hong Kong,” IEEE Transactions on Visualization and Computer Graphics, vol. 13, no. 6, 2007, pp. 1408-1415.
[13]K. Gu, J. Qiao and W. Lin, “Recurrent Air Quality Predictor Based on Meteorology and Pollution-Related Factors,” IEEE Transactions on Industrial Informatics, vol. 14, no. 9, 2018, pp. 3946-3955.
[14]A. Boubrima, W. Bechkit and H. Rivano, “Optimal WSN Deployment Models for Air Pollution Monitoring,” IEEE Transaction Wireless Communications, vol. 16, no. 5, 2017, pp. 2723-2735.
[15]A. R. Al-Ali, I. Zualkernan and F. Aloul, “A Mobile GPRS-Sensors Array for Air Pollution Monitoring,” IEEE Sensors Journal, vol. 10, 2010, pp. 1666-1671.
[16]C. Xiaojun, L. Xianpeng and X. Peng, “IoT-Based Air Pollution Monitoring and Forecasting System,” International Conference on Computer and Computational Sciences, 2015, pp. 257-260.
[17]N. Dam, A. Ricketts, B. Catlett and J. Henriques, “Wearable Sensors for Analyzing Personal Exposure to Air Pollution,” System and Information Engineering Design Symposium, 2017, pp. 1-4.
[18]D. Qin, J. Yu, G. Zou, R. Yong and Q. Zhao, “A Novel Combined Prediction Scheme Based on CNN and LSTM for Urban PM2.5 Concentration,” IEEE Access, vol. 7, 2019, pp. 20050-20059.
[19]S. Duangsuwan, A. Takarn, R. Nujankaew and P. Jamjareegulgarn, “A Study of Air Pollution Smart Sensors LPWAN via NB-IoT for Thailand Smart Cities 4.0,” IEEE International Conference on Knowledge and Smart Technology, 2018.
[20]B. Gates, N. Myhrvold and P. Rinearson, “The Road Ahead,” Viking Penguin, 1995.
[21]M. Dholu and K. Ghodinde, “Internet of Things for Precision Agriculture Application,” International Conference on Trends in Electronics and Informatics, 2018.
[22]T. C. Chan, “The Development of Smart Manufacturing and Cases Study in Taiwan,” IEEE International Conference on Advanced Manufacturing, 2018.
[23]鄭福炯，物聯網技術理論與實作；第二版，新北：全華圖書，2019。
[24]Computer Network Types Wikipedia, https://en.wikipedia.org/wiki/Computer_network