以機器學習對移動電話資料進行運具種類的智慧判讀;Inferring Transportation Modes from Mobile Phone Data Using Machine Learning Methods

NCUIR > Engineering College > Department of Civil Engineering > Research Project > Item 987654321/84460

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/84460

Title:	以機器學習對移動電話資料進行運具種類的智慧判讀;Inferring Transportation Modes from Mobile Phone Data Using Machine Learning Methods
Authors:	陳惠國
Contributors:	土木工程學系
Keywords:	行動電話資料;飄移現象;運具判斷;支持向量機;深度學習;mobile phone data;oscillation phenomenon;mode inference;vector support machine;deep learning
Date:	2020-12-08
Issue Date:	2020-12-09 09:22:31 (UTC+8)
Publisher:	科技部
Abstract:	運輸規劃旅運需求預測中最重要的幾個步驟分別為旅次產生、旅次分派、運具選擇以及交通量指派。在過去都是利用家戶或是路邊訪問的問卷調查方式獲得運輸規劃所需的資料，但透過問卷調查的方式通常面臨了(1)耗費大量人力、(2)高拒訪率以及(3)因受訪者記憶不完整而造成錯誤填答。近年來雖然已經嘗試利用GPS資料取代過去的調查方法，但GPS資料除了不易獲得外還容易受到建築物的遮蔽效應造成定位不準確，因此不適合將GPS資料應用於大型路網中。而行動電話資料已成為運輸規劃另一種資料蒐集方式，它無須新增額外設備即可自動且有效率的紀錄使用者的時空資料。因此，獲得行動電話資料的成本很低，甚至可以忽略不計。在本研究當中，我們擬採用支持向量機(SVM)與深度神經網路(DNN)兩種監督式機器學習的方法探討在何種特徵(旅行時間、蹤跡的發生時間、兩蹤跡間(traces)的速度、旅程中最大的速度以及平均速度)、時間區段(尖峰時段、離峰時段以及全天)、運具路線組合(公共汽車路線、汽車行走於公共汽車路線、與汽車行走於非公共汽車路線)以及訓練的方法等因素以及不同因素組合將會如何影響運具判斷的準確性。本研究擬以利用行動電話資料進行運具判斷為例，將計畫分成兩年執行。第一年的工作項目包括：(1) 文獻回顧； (2) 研究架構建立； (3) 大規模實地資料調查以及連繫電信公司撈取所調查之信令資料； (4) 移動手機現地資料(sighting data)前處理； (5) 支持向量機(SVM)結果分析；(6) 期中報告。第二年的工作項目包括：(1) 文獻回顧； (2) 研究架構建立； (3) 將進行大型路網之車輛/巴士路線的分群、並收集樣本巴士路線的票證資料； (4) 深度神經網路(DNN) 結果分析； (5) 期末報告。本計畫預估可產出一至兩篇國際期刊論文。 ;Traffic data – consisting of activity location, origin-destination pair, mode choice and traffic assignment – are essential in transportation planning, or more specifically, travel demand forecasting. Collecting such data via a questionnaire survey, like the home or roadside interview, have long been adopted, but are usually (1) labor intensive, (2) faced with high refusal rates of respondents, and (3) relatively inaccurate due to fade-away memory. Attempts have been made to use GPS data, but GPS data are not readily available and their levels of accuracy are apt to be affected by the shielding effect due to high-rise buildings and obstacles and, hence, are not suitable to be applied in a large transportation network. Mobile phone data, emerging as a vivid data collection method for transportation planning, can automatically and effectively record transportation planning data in time-space dimension without having to add new devices. Thus, the extra cost to retrieve this phone data is small or even negligible. For this study, we will adopt two supervised machine learning methods – support vector machine (SVM) and deep neural network (DNN) – to investigate how modal features (travel times, starting time of traces, traversal speeds between traces, maximum speeds, and average speeds), time of day (peak hours, off-peak hours, whole day), route combinations (bus route, vehicle traversing a bus route, vehicle traversing non-bus routes), and training methods, i.e., support vector machine (SVM) or deep neural network (DNN) affect accuracy in inferring transportation modes (either bus or vehicle).This study takes mode inference as an example to explore the usefulness of mobile phone data in the area of transportation planning. This proposal will be conducted in two consecutive years. In the first year, the major work to be carried out covers: (1) literature review; (2) research framework; (3) field data collection; (4) preprocessing of mobile phone data (sighting data); (5) SVM result analysis; (6) interim project report. In the second year, the major work to be carried out includes: (1) literature review; (2) research framework; (3) classification of vehicle/bus routes for a practical transportation network and collection of smart card data for bus routes; (4) DNN result analysis; (5) final project report. When the two-year project is completed, the results thus obtained are expected to come up with one or two papers to be published in international journals.
Relation:	財團法人國家實驗研究院科技政策研究與資訊中心
Appears in Collections:	[Department of Civil Engineering] Research Project

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	159	View/Open

社群 sharing

Loading...