請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/87011
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 周俊廷 | zh_TW |
dc.contributor.advisor | Chun-Ting Chou | en |
dc.contributor.author | 邱正鈞 | zh_TW |
dc.contributor.author | Cheng-Chun Chiu | en |
dc.date.accessioned | 2023-05-02T17:24:18Z | - |
dc.date.available | 2023-11-09 | - |
dc.date.copyright | 2023-05-02 | - |
dc.date.issued | 2023 | - |
dc.date.submitted | 2023-01-14 | - |
dc.identifier.citation | T. Friedrich, M.S. Krejca, R. Rothenberger, T. Arndt, D. Hafner, T. Kellermeier, S. Krogmann, and A. Razmjou. Routing for on-street parking search using probabilistic data. In AI Communications, volume 32, pages 113–124. 2019.
Q. Chen, A. Conway, and J. Cheng. Parking for residential delivery in new york city: Rgulations and behavior. In Elsevier, volume 54, pages 53–60. 2017. P. E. Carnelli, J. Yeh, M. Sooriyabandara, and A. Khan. Parkus: A novel vehicle parking detection system. Proc. 31st AAAI Conf. Artif. Intell., pages pp. 4650–4656. 2017. S. Gomari, C. Knoth, and C. Antoniou. Cluster analysis of parking behaviour: A case study in munich. In Transportation Research Procedia, volume 52, pages 485–492. 2021. Zoom levels and tile grid. https://learn.microsoft.com/en-us/azure/azure-maps/zoom-levels-and-tile-grid?tabs=csharp. S. Reed, A. Melissa Campbell, and B. W. Thomas. Does parking matter? the impact of search time for parking on last-mile delivery optimization. In Working paper, arXiv:2107.06788v2. 2022. S. Reed, A. Melissa Campbell, and B. W Thomas. The value of autonomous vehicles for last-mile deliveries in urban environments. In Management Science, volume 68(1), pages 280–299. 2021. A. Martinez-Sykora, F. McLeod, C. Lamas-Fernandez, T. Bektas, T. Cherrett, and J. Allen. Optimised solutions to the last-mile delivery problem in london using a combination of walking and driving. In Annals of Operations Research, volume 295(2), pages 645–693. 2020. T. Ba T Nguyen, T. Bektas, T. J Cherrett, F. N McLeod, J. Allen, O. Bates, M. Piotrowska, M. Piecyk, A. Friday, and S. Wise. Optimising parcel deliveries in London using dual-mode routing. In Journal of the Operational Research Society, volume 70(6), pages 998–1010. 2020. W. Zou, X. Wang, A. Conway, and Q. Chen. Empirical analysis of delivery vehicle on-street parking pattern in manhattan area. In J. Urban Plan. Dev, editor, Journal of Urban Planning and Development, volume 142(2): 04015017. 2016. Object tracking implemented with yolov4, deepsort, and tensorflow. https://pythonrepo.com/repo/theAIGuysCode-yolov4-deepsort-python-deep-learning. Google Maps Platform Distance Matrix API. https://developers.google.com/maps/documentation/distance-matrix/overview. Google Maps Platform Geocoding API. https://developers.google.com/maps/documentation/geocoding/overview. | - |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/87011 | - |
dc.description.abstract | 隨著全球疫情的肆虐以及網路的發達,無接觸採買以及上網購物的需求增加,如何進行有效率得配送變得愈發重要。在這樣的趨勢下,物流業雇用越來越多沒有太多經驗的人來作為職業司機將變得無可避免,因此,如何有效地培養新手司機將貨物有效率地配送到每一位顧客手上變成一項非常重要的課題。
本篇論文旨在藉由學習老手司機的送貨策略來解決上述問題,以減少新手司機的訓練時間以及金錢成本。我們採用深度神經網路(deep neural network, DNN)以及基於規則(Rule-based)的方法來判別老手司機送貨路線中的停車地點,並將這些停車地點與當日的配送地址結合,來得知老手司機如何在送貨的過程中,根據送貨地址決定停車位置。 結果顯示在判斷送貨停車點的準確度包含precision, accuracy及recall分別可以達到91%、92%以及94%以上,而送貨地址配對停車點的正確率亦可超過95%。在測試資料中,113個建議停車點中與對照組方法相比,我們可減少6筆送貨地址與建議停車點距離超過30公尺之案例。 | zh_TW |
dc.description.abstract | The increasing importance of improving logistic efficiency has become critical in recent times due to the global impact of the epidemic and the growth of the internet. As a result, it has become necessary to hire more drivers, including those with little experience as professional drivers, in order to effectively deliver packages to customers. The training of these inexperienced drivers poses a significant challenge.
This thesis aims to address this issue by learning from the delivery strategies of experienced drivers to reduce training time and related expenses. We adopt deep neural networks (DNN) and rule-based methods to identify the delivery parking spots of experienced drivers along the delivery route and pair these delivery parking spots with the delivery addresses of the packages. Our results show that the precision, accuracy, and recall of the detection of delivery parking spots are all above 91%, 92%, and 94%, respectively. The accuracy of pairing results of delivery packages and delivery parking spots can exceed 95%. Compared with the baseline method, our proposed method can reduce the number of cases where the distance between the delivery address and the recommended parking spot is more than 30 meters by 6. | en |
dc.description.provenance | Submitted by admin ntu (admin@lib.ntu.edu.tw) on 2023-05-02T17:24:18Z No. of bitstreams: 0 | en |
dc.description.provenance | Made available in DSpace on 2023-05-02T17:24:18Z (GMT). No. of bitstreams: 0 | en |
dc.description.tableofcontents | Acknowledgements i
摘要 ii Abstract iii Contents v List of Figures viii List of Tables x Chapter 1 INTRODUCTION 1 1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2.1 Identification of “parkings” during the delivery . . . . . . . . . . . 4 1.2.2 Identification of the purpose of the detected parking . . . . . . . . . 5 1.2.3 Associate of the parking for delivery with the packages . . . . . . . 6 1.2.4 Summary of Related Work . . . . . . . . . . . . . . . . . . . . . . 8 1.2.5 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.2.6 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.2.7 Organization of the Thesis . . . . . . . . . . . . . . . . . . . . . . 9 Chapter 2 SYSTEM SETTINGS 10 2.1 Description of Input Data . . . . . . . . . . . . . . . . . . . . . . . . 10 2.1.1 Driving Videos . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.1.2 GPS Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.1.3 Acc Status Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.1.4 Delivery Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.2 Key Idea of the Proposed Solution . . . . . . . . . . . . . . . . . . . 12 Chapter 3 PROPOSED SOLUTIONS 14 3.1 Definition of Notation . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.2 Pipeline of the Proposed Solutions . . . . . . . . . . . . . . . . . . . 18 3.3 Step 1: Finding delivery parking spots . . . . . . . . . . . . . . . . . 19 3.3.1 Step 1-0: Moving detection . . . . . . . . . . . . . . . . . . . . . . 20 3.3.2 Step 1-1: Check acc status . . . . . . . . . . . . . . . . . . . . . . 21 3.3.3 Step 1-2: vehicle in front detection . . . . . . . . . . . . . . . . . . 23 3.3.4 Step 1-3: Traffic light detection . . . . . . . . . . . . . . . . . . . . 28 3.4 Step 2: Pair parking spots with addresses of delivery addresses . . . . 31 3.4.1 Step 2-1: Calculate the GPS distance from all eligible parking spots to the delivery address . . . . . . . . . . . . . . . . . . . . . . . . 32 3.4.2 Step 2-2: Calculate the walking distance from the first three closest parking spots to the delivery address . . . . . . . . . . . . . . . . . 33 3.4.3 Step 2-3: Improve the pairing rate of parking spots and delivery addresses in the same alley . . . . . . . . . . . . . . . . . . . . . . . 35 3.4.4 Step 2-4: Choose the best parking spot under the “close enough” criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.5 Step 3: Recommend parking spots for the new delivery packages . . 36 3.5.1 Step 3-1: Map the address of the new package to the nearest delivery address in the past . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 3.5.2 Step 3-2: Use DBSCAN to cluster the parking spots that the nearest past delivery address paired with . . . . . . . . . . . . . . . . . . . 38 3.5.3 Step 3-3: Select a point in the cluster with the most elements as the recommended parking spot . . . . . . . . . . . . . . . . . . . . . . 40 Chapter 4 PERFORMANCE EVALUATION 41 4.1 Experimental Datasets & Settings . . . . . . . . . . . . . . . . . . . 41 4.1.1 Experimental Datasets . . . . . . . . . . . . . . . . . . . . . . . . 41 4.1.2 The Procedure of Labeling Ground Truth Data . . . . . . . . . . . . 43 4.1.3 Performance Metrics . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.1.3.1 Stage 1 . . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.1.3.2 Stage 2 . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.1.3.3 Stage 3 . . . . . . . . . . . . . . . . . . . . . . . . . . 48 4.2 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4.2.1 Performance of the Delivery Parking Spot Detection . . . . . . . . . 49 4.2.2 Comparison of Proposed Method and Baseline Method . . . . . . . 63 4.2.2.1 Pairing Results . . . . . . . . . . . . . . . . . . . . . . 63 4.2.2.2 Recommendation Results . . . . . . . . . . . . . . . . 70 Chapter 5 CONCLUSIONS 76 References 78 | - |
dc.language.iso | en | - |
dc.title | 老手駕駛的數位孿生: 以資料驅動的最後一哩配送解決方案 | zh_TW |
dc.title | Digital Twin of Experienced Drivers for Last-Mile Delivery — A Data-Driven Approach | en |
dc.type | Thesis | - |
dc.date.schoolyear | 111-1 | - |
dc.description.degree | 碩士 | - |
dc.contributor.oralexamcommittee | 謝宏昀;魏宏宇;逄愛君 | zh_TW |
dc.contributor.oralexamcommittee | Hung-Yun Hsieh;Hung-Yu Wei;Ai-Chun Pang | en |
dc.subject.keyword | 機器學習,貨物配送,停車點分析,經驗學習,影像辨識, | zh_TW |
dc.subject.keyword | machine learning,delivery of packages,parking spot analysis,experiential learning,image recognition, | en |
dc.relation.page | 79 | - |
dc.identifier.doi | 10.6342/NTU202300096 | - |
dc.rights.note | 同意授權(全球公開) | - |
dc.date.accepted | 2023-01-16 | - |
dc.contributor.author-college | 電機資訊學院 | - |
dc.contributor.author-dept | 電信工程學研究所 | - |
顯示於系所單位: | 電信工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-111-1.pdf | 13.39 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。