請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70264完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 李允中(Yeun-Chung Lee) | |
| dc.contributor.author | Chao-Liang Hsieh | en |
| dc.contributor.author | 謝兆糧 | zh_TW |
| dc.date.accessioned | 2021-06-17T04:24:58Z | - |
| dc.date.available | 2021-08-18 | |
| dc.date.copyright | 2018-08-18 | |
| dc.date.issued | 2018 | |
| dc.date.submitted | 2018-08-15 | |
| dc.identifier.citation | 陳建旭、劉韋廷、廖敏志、王慶雄、蔡益智、林和志。2011。分析鋪面坑洞產生原因與建議維護方法。臺灣公路工程 第 37 卷第 10 期。
行政院交通部。2017。105年道路長度及橋梁座數概況。台北：行政院交通部。網址：https://www.motc.gov.tw/ch/home.jsp?id=64&parentpath=0,6。上網日期：2017-07-14。 行政院交通部。2016。道路交通事故。台北：行政院交通部。網址：http://stat.motc.gov.tw/mocdb/stmain.jsp?sys=100&funid=b3303。上網日期：2016-08-12。 Aizerman, M. A. 1964. Theoretical foundations of the potential function method in pattern recognition learning. Automation and remote control 25: 821-837. Aminian, K., P. Robert, E. Jequier, and Y. Schutz. 1995. Estimation of speed and incline of walking using neural network. IEEE Transactions on Instrumentation and Measurement 44(3): 743-746. Aminian, K., P. Robert, E. E. Buchser, B. Rutschmann, D. Hayoz, and M. Depairon. 1999. Physical activity monitoring based on accelerometry: validation and comparison with video observation. Medical & biological engineering & computing 37(3): 304-308. Bao, L., and S. S. Intille. 2004. Activity recognition from user-annotated acceleration data. In 'International Pervasive Computing Conference', 1-17, Linz/Vienna, Austria. Boser, B. E., I. M. Guyon, and V. N. Vapnik. 1992. A training algorithm for optimal margin classifiers. In 'Proc. 5th ACM Computational learning theory workshop', 144-152, Pittsburgh, PA, USA: ACM. Brisimi, T. S., C. G. Cassandras, C. Osgood, I. C. Paschalidis, and Y. Zhang. 2016. Sensing and Classifying Roadway Obstacles in Smart Cities: The Street Bump System. IEEE Access 4: 1301-1312. Chambers, G. S., S. Venkatesh, G. A. West, and H. H. Bui. 2002. Hierarchical recognition of intentional human gestures for sports video annotation. In 'Proc. 16th IEEE Pattern Recognition Conference', 1082-1085, Quebec, Canada: IEEE. Chan, K. P., and A. W. C. Fu. 1999. Efficient time series matching by wavelets. In 'Proc. 15th IEEE Data Engineering Conference', 126-133, Sydney, Australia: IEEE. Cortes, C., and V. Vapnik. 1995. Support-vector networks. Machine learning: 20(3), 273-297. Denver Business Journal. 2007. Money pits. Available at: https://www.bizjournals.com/denver/stories/2007/04/02/story1.html?b=1175486400%25255E1438887. Eriksson, J., L. Girod, B. Hull, R. Newton, S. Madden, and H. Balakrishnan. 2008. The pothole patrol: using a mobile sensor network for road surface monitoring. In 'Proc. 6th ACM Mobile systems, applications, and services conference', 29-39, Breckenridge, CO, USA: ACM. Farringdon, J., A. J. Moore, N. Tilbury, J. Church, and P. D. Biemond. 1999. Wearable sensor badge and sensor jacket for context awareness. In '3th IEEE Wearable Computers Symposium', 107-113, San Francisco, USA: IEEE. Figo, D., P. C. Diniz, D. R. Ferreira, and J. M. Cardoso. 2010. Preprocessing techniques for context recognition from accelerometer data. Personal and Ubiquitous Computing 14(7): 645-662. González, L. C., R. Moreno, H. J. Escalante, F. Martínez, and M. R. Carlos. 2017. Learning roadway surface disruption patterns using the bag of words representation. IEEE Transactions on Intelligent Transportation Systems 18(11): 2916-2928. Hastie, T., R. Tibshirani, and J. Friedman. 2009. Overview of supervised learning. In The elements of statistical learning on Springer New York: 9-41. Hecht-Nielsen, R. 1987. Counterpropagation networks. Applied optics 26(23): 4979-4984. Hecht-Nielsen, R. 1988. Applications of counterpropagation networks. Neural networks 1(2): 131-139. Hecht-Nielsen, R. 1992. Theory of the backpropagation neural network. In Neural networks for perception: 65-93. Hsu, C. W., C. C. Chang, and C. J. Lin. 2003. A practical guide to support vector classification. Jeong, D. U., S. J. Kim, and W. Y. Chung. 2007. Classification of posture and movement using a 3-axis accelerometer. In 'IEEE Convergence Information Technology Conference', 837-844, Gyeongju, South Korea: IEEE. Jetpatcher. 2018. ASPHALT ROAD – FAILURES. Available at: http://jetpatcher.co.za/. Accessed 15 April 2018. Kawahara, Y., H. Kurasawa, and H. Morikawa. 2007. Recognizing user context using mobile handsets with acceleration sensors. In 'IEEE Portable Information Devices Conference', 1-5, Orlando, USA: IEEE. Kotsiantis, S. B., D. Kanellopoulos, and P. E. Pintelas. 2006. Data preprocessing for supervised leaning. International Journal of Computer Science 1(2): 111-117. Magoosh. 2017. Data Science. Available at: https://magoosh.com/data-science/k-fold-cross-validation/. Accessed 8 December 2017. Martens, W. L. 1992. The fast time frequency transform (FTFT): a novel on-line approach to the instantaneous spectrum. In '14th IEEE Engineering in Medicine and Biology Society conference', 2594-2595. Paris, France: IEEE. Olvera-López, J. A., J. A.Carrasco-Ochoa, J. F. Martínez-Trinidad, and J. Kittler. 2010. A review of instance selection methods. Artificial Intelligence Review 34(2): 133-143. Rahm, E., and H. H. Do. 2000. Data cleaning: Problems and current approaches. IEEE Data Eng. Bull. 23(4): 3-13. Randell, C., and H. Muller. 2000. Context awareness by analysing accelerometer data. In ' 4th IEEE Wearable Computers Symposium', 175-176, Atlanta, USA: IEEE. Scornet, E. 2016. Random forests and kernel methods. IEEE Transactions on Information Theory 62(3): 1485-1500. Seraj, F., B. J. van der Zwaag, A. Dilo, T. Luarasi, and P. Havinga. 2014. RoADS: A road pavement monitoring system for anomaly detection using smart phones. In Big data analytics in the social and ubiquitous context on Springer, Cham: 128-146. Sklearn. 2017. sklearn.metrics.accuracy_score. Available at: http://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_fscore_support.html. Accessed 10 August 2017. Sklearn. 2017. sklearn.metrics.precision_recall_fscore_support. Available at: http://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_fscore_support.html. Accessed 10 August 2017. Towards Data Science. 2017. A Gentle Introduction To Neural Networks Series — Part 1. Available at: https://towardsdatascience.com/a-gentle-introduction-to-neural-networks-series-part-1-2b90b87795bc. Accessed 4 March 2018. Van Laerhoven, K., K. A. Aidoo, and S. Lowette. 2001. Real-time analysis of data from many sensors with neural networks. In 'Proc. 5th IEEE Wearable Computers Symposium', 115-122, Zurich, Switzerland: IEEE. Vapnik, V. 1998. The support vector method of function estimation. Nonlinear modeling: Advanced black-box techniques, 55, 86. Vapnik, V. N. and Vapnik, V. 1998. Statistical learning theory (Vol. 1). New York: Wiley. Vapnik, V. 2013. The nature of statistical learning theory. Springer science & business media. World Economic Forum. 2016. The number of cars worldwide is set to double by 2040. Bernstein. Available at: https://www.weforum.org/agenda/2016/04/the-number-of-cars-worldwide-is-set-to-double-by-2040. Accessed 22 April 2017. World Highways. 2015. The importance of road maintenance. Available at: http://www.worldhighways.com/categories/maintenance-utility/features/the-importance-of-road-maintenance/. Wu, S. 2013. A review on coarse warranty data and analysis. Reliability Engineering & System Safety 114: 1-11. | |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70264 | - |
| dc.description.abstract | 道路連接了建築、村落、甚至是城市，在生活中扮演著相當重要的角色，其衍生出的價值是相當可觀的，無疑是社會最重要的基礎設施之一。在台灣，道路總長度為43365公里，整體道路網絡串連了台灣的經貿、人流、交通，降低整個台灣的空間尺度，也因此縮短來往各地時間。若道路品質不好，一條路上有許多坑洞或下陷道路等情況，會造成乘坐不適、駕駛和乘客安全疑慮、車輛懸吊系統磨損、以及交通意外等問題。因此，道路品質與維護修繕顯得極為重要。目前台灣道路修繕維護方式主要是以工程車巡視、民眾回報和定期修繕為主，需要花費大量人力和時間，才能正確地找到需要維修的路段。為了維護道路品質與改善政府修繕效率，本計畫提出了一套基於物聯網技術之道路表面不規則偵測系統。此系統之前端感測節點搭載震動感測器、GPS模組、4G傳輸模組，當震動振幅超過設定閾值時，會連續量測一段時間，記錄此時的震動波形、經緯度和車速，透過4G傳輸模組，回傳到後端資料庫。並於後端運算系統將收集到的各種道路表面型態(如：平坦道路、坑洞、人孔蓋、下陷道路等)之波形進行數據分析，並透過機器學習方法來進行道路表面型態辨識，並可在Google Map上顯示其辨識結果，進而提供民眾和政府單位參考，政府單位可依照道路之嚴重程度選擇優先修繕路段。如此一來，便可大為減少原先檢視道路表面狀況所需之人力與時間成本，更可提升道路修繕之效率。 | zh_TW |
| dc.description.abstract | Roads connecting buildings, villages and even cities play a very important role in our life. The values derived from them are considerable, and they are undoubtedly one of the most important infrastructures in society. In Taiwan, the total length of the roads is 43,365 kilometers. The overall road network links Taiwan's economy, trade, people, and transportation, reducing the spatial scale of Taiwan as a whole, and shortening the travel time to and from all places. If the road quality is not good, there are many potholes or roads that are sloping down the road on one road. This can cause problems such as uncomfortable rides, driving and passenger safety concerns, vehicle suspension system wear, and traffic accidents. Therefore, road quality and maintenance repairs are extremely important. At present, the maintenance of roads in Taiwan is mainly based on inspections of construction vehicles, returns from the public, and regular repairs. It takes a lot of manpower and time to find the correct road sections that need maintenance. In order to maintain road quality and improve the efficiency of government repairs, an anomaly detection system for roadway surface monitoring based on IoT and machine learning technologies is proposed in this study. The front-end sensing node of this system is equipped with a vibration sensor, a GPS module, and a 4G transmission module. When the vibration amplitude exceeds the set threshold, continuous measurement is performed for a period of time to record the vibration waveform, latitude and longitude, and vehicle speed at the time through 4G transmission module, back to the back-end database. In addition, the back-end computing system analyzes the waveforms of various road surface types (such as regular roads, potholes, manholes, and depressions) and uses machine learning methods to identify road surface types. And these classification results can be displayed on Google Map, and then provide reference for the public and government agencies. Government agencies can choose to repair road sections according to the severity of the road. As a result, the manpower and time costs which are required to examine the surface conditions of the roads can be greatly reduced, and the efficiency of road repairs can be improved. | en |
| dc.description.provenance | Made available in DSpace on 2021-06-17T04:24:58Z (GMT). No. of bitstreams: 1 ntu-107-R05631025-1.pdf: 6004986 bytes, checksum: ba6301f05d76654991769c9955c72ea4 (MD5) Previous issue date: 2018 | en |
| dc.description.tableofcontents | Table of Contents
誌謝 i 摘要 iii Abstract v Table of Contents vii List of Figures xi List of Tables xv Chapter 1 Introduction 1 1.1 Background 1 1.2 Motivation and purpose 3 1.3 Thesis organization 4 Chapter 2 Literature Review 5 2.1 Roadway surface monitoring system 5 2.1.1 Pothole and anomaly detection system for road surface monitoring 6 2.1.2 Sensing and classifying roadway obstacles in smart cities: the street bump system 8 2.1.3 Road pavement anomaly detection system 10 2.1.4 Bag of words representation 11 2.2 Data pre-processing for constructing an anomaly detection model 13 2.2.1 Data cleaning 14 2.2.2 Instance selection and feature extraction 15 2.2.3 Data normalization 17 2.2.4 Feature extraction 19 2.3 Overview of different machine learning methods 20 2.3.1 Support vector machine 20 2.3.2 Random forests 21 2.3.3 AdaBoost with stumps 22 2.3.4 Artificial neural networks 22 2.3.5 Machine learning methods for roadway surface classification 24 Chapter 3 Methods 27 3.1 Anomalies detection system for different roadway surface conditions 27 3.2 Setting-up and data collection 28 3.3 Pre-processing techniques utilized in this thesis 31 3.3.1 Time domain 32 3.3.2 Frequency domain 34 3.4 Support vector machine for roadway surface conditions classification 35 3.4.1 Linear SVM 35 3.4.2 Nonlinear classification 38 3.4.3 SVM classifier 40 3.4.4 Properties 42 3.5 Index of model performance 43 3.5.1 K-fold cross validation 43 3.5.2 Accuracy score 45 3.5.3 Precision, recall, and f1-Score 45 Chapter 4 Research Results 47 4.1 Dataset 48 4.2 Time domain and frequency domain analysis 51 4.2.1 Definition of attributes utilized in this study 52 4.2.2 Time and frequency domain analyses for different types of roadway surfaces 55 4.3 Classification model for roadway surfaces 66 4.3.1 Splitting the labeled data set into the training, validation and testing sets 66 4.3.2 Obtaining the best parameter in the training stage 70 4.3.3 Testing stage where the model performance is evaluated 81 4.3.4 Testing the classification model with the unlabeled dataset 86 4.4 Classification results shown on google maps 87 Chapter 5 Conclusions and Future Work 89 References 91 | |
| dc.language.iso | zh-TW | |
| dc.subject | 道路表面不規則偵測系統 | zh_TW |
| dc.subject | 物聯網技術 | zh_TW |
| dc.subject | 機器學習 | zh_TW |
| dc.subject | Internet of Things (IoTs) | en |
| dc.subject | Anomaly detection system | en |
| dc.subject | roadway surface monitoring | en |
| dc.subject | machine learning | en |
| dc.title | 基於物聯網與機器學習技術之不規則道路表面偵測系統 | zh_TW |
| dc.title | An anomaly detection system for roadway surface monitoring based on IoT and machine learning technologies | en |
| dc.type | Thesis | |
| dc.date.schoolyear | 106-2 | |
| dc.description.degree | 碩士 | |
| dc.contributor.coadvisor | 江昭皚(Joe-Air Jiang) | |
| dc.contributor.oralexamcommittee | 溫在弘(Tzai-Hung Wen),曾傳蘆(Chwan-Lu Tseng) | |
| dc.subject.keyword | 道路表面不規則偵測系統,物聯網技術,機器學習, | zh_TW |
| dc.subject.keyword | Anomaly detection system,roadway surface monitoring,Internet of Things (IoTs),machine learning, | en |
| dc.relation.page | 94 | |
| dc.identifier.doi | 10.6342/NTU201803338 | |
| dc.rights.note | 有償授權 | |
| dc.date.accepted | 2018-08-15 | |
| dc.contributor.author-college | 生物資源暨農學院 | zh_TW |
| dc.contributor.author-dept | 生物產業機電工程學研究所 | zh_TW |
| 顯示於系所單位： | 生物機電工程學系 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-107-1.pdf 未授權公開取用 | 5.86 MB | Adobe PDF |
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。