6. Discussion 77
6.8. Learning Reflections
Since this was the first project working with real-life data for the both of us, we have made great new experiences on which we will reflect in this section. Compared to projects and assignments in our courses, where data sets usually are already preprocessed and suitable for the use of various machine learning models, the data that we have been provided with for this project was originally not designed for a data-driven approach. It was hard to estimate the workload of the data preparation and preprocessing, which we underestimated several times within this project. In the end, around 80% of the time that we spent on coding went into this step. Furthermore, we had to learn to make logical assumptions, for example to fill missing data points.
However, we have also gained new competencies. Theoretical concepts and models that we have studied in our courses in previous semesters have now been useful to determine our research goals and we could apply them in practice. Thereby, both of us have significantly improved our coding skills and increased the practical experience of working on data analytics.
We have also mastered organizational challenges given the fact that various stakeholders with different interests have been involved in the project.
In this thesis we exploredhow artificial intelligence and machine learning as data-driven technologies can improve the process of road maintenance towards data-driven decision making. To answer the research question comprehensively, we combined road maintenance data from three incompatible systems from our case study, the City of Copenhagen, and explored different possibilities to apply machine learning models on it. Furthermore, we explored the topic from a theoretical angle through a literature and case study analysis.
With the help of various visualizations we shared first insights into the data that have not existed in such a form previously. Furthermore, we created a pavement condition index (PCI) to compare the conditions across different roads and assigned all road segments a score between 0 and 100. We then applied a linear regression and XGBoost model on the data to predict above-named PCI, where our best model, an XGBoost model with optimized hyperparameters, reached an adjustedR2score of 0.9011 with an RMSE of 11.74. Both models have the ability to reveal the importance of the different input features, which led us conclude that the damages alligator cracks, large cracks and rutting have the biggest influence on road deterioration.
In the second part of our analysis, we predict acute damages in the form of potholes one year ahead for each road segment. The received scores show that including all available data does improve the performance of the predictions compared to our baseline by 30%. However, with an adjusted R2 score of 0.3977 for our best model, the well-appearing RMSE score of 1.32 can be misleading and should not be put into production in its current maturity degree.
The theoretical part of our thesis focused on exploring different technologies that serve as an inspiration to improve the future road maintenance process for the City of Copenhagen.
Based on our findings, we created two scenarios with technologies that are either already
in large-scale use in other regions or deliver - according to the academic literature - highly accurate and cost-effective results.
The successful use of artificial intelligence and machine learning to optimize the process of road maintenance has been proven both in theory and practice. Current road maintenance processes at the City of Copenhagen would benefit extremely from those technologies in regard to better quality and overview of assessments, as well as an improved allocation of available budgets. At the present time, the available systems and the data that they contain about road assessments are not designed to be used for predictive maintenance activities, which makes such data analysis laborious and inaccurate. We combined all our insights gathered during the process to identify five recommendations for the City of Copenhagen.
All of the recommendation include a data-centric perspective to help the municipality to move towards data-driven decision making for predictive road maintenance. Furthermore, we outline future work and potential deliverables from a practical and research approach.
In this master thesis we contributed to the research of data-driven decision-making for predictive road maintenance in two points. First, we proposed a new approach to create a pavement condition index (PCI) based on the estimated remaining lifetime of a road. Secondly, we came up with a proposal for future data collection, of which requirements need to be met in order to use road maintenance data for predictions and towards data-driven decision making.
We conclude that the current state of data maturity at the City of Copenhagen needs to improve to effectively leverage data-driven technologies for the process of road maintenance.
However, we showed the potential which can be further exploited in future work.
2.1. Example of a linear regression . . . 14 2.2. Example of a decision tree . . . 15 3.1. Overview of Systems and Stakeholders handling Road Maintenance at the TMF 22 4.1. Saunders Research Onion . . . 32 4.2. CRISP-DM Lifecycle . . . 34 4.3. Example filling missing values . . . 40 4.4. Example adding "construktionsname" . . . 41 4.5. Mapping PUMA task to RoSy Road Segment . . . 46 4.6. Time Span of utilized data for Acute Damage Prediction . . . 48 5.1. All streets that are included in the RoSy data set . . . 55 5.2. Pothole tasks from PUMA data set . . . 55 5.3. Reported potholes from Giv et praj data set . . . 56 5.4. Potholes from PUMA and Giv et praj . . . 56 5.5. Data exploration overview after preprocessing . . . 57 5.6. Descriptive values of all normalized damages from RoSy . . . 58 5.7. PCI of all data entries of the road segmentAmagerbrogade-6-L0over years with
changing values ofFromChainageandToChainage . . . 59 5.8. Overview about road conditions by means of PCI . . . 60 5.9. True Values vs. Predictions (PCI Predictions) . . . 62 5.10. Residual Distribution (PCI Predictions) . . . 62 5.11. Feature Importance Linear Regression . . . 63 5.12. Feature Importance XGBoost (Baseline) . . . 64
5.13. Feature Importance XGBoost (Optimized) . . . 65 5.14. True Values vs. Predictions (Time Series Prediction) . . . 66 5.15. Residual Distribution (Time Series Prediction) . . . 66 6.1. Exemplary Dashboard . . . 89
3.1. Types of road damages . . . 25 3.2. Damage limits that reduce RSL . . . 26 4.1. Academic literature search used keywords and results . . . 52 4.2. Case Study search keywords and results . . . 52 5.1. Results of PCI Predictions . . . 61 5.2. Results of Acute Damage Predictions . . . 65 5.3. Overview of identified procedures for automated assessments of road quality 68
AI Artifical Intelligence. 11 EY Ernst & Young. 3, 52
FWD Falling Weight Deflectometer. 8, 9, 71 GPR Ground Penetration Radar. 9, 71, 72
HWD Heavy Falling Weight Deflectometer. 9, 71, 72
ICT Information and Communication Technology. 1, 6 IoT Internet of Things. 1
IRI International Roughness Index. 8, 79 ML Machine Learning. 11
PCI Pavement Condition Index. 2, 8, 54, 91 PMS Pavement Management System. 8, 9, 22 PSI Present Serviceability Index. 8
PSR Present Serviceability Ratio. 8
PUMA Platform til Understøttelse af Mobile Arbejdsgange. 23, 36, 39 RoSy RoSy cRoad Asset Management System. 22, 36, 38
RSL Remaining Service Life. 8, 72 RSME Root Mean Squared Error. 18
TMF Teknik- og Miljøforvaltningen (Technical and Environmental Management). 21, 36, 52
[1] P. Gamage. “New development: Leveraging ‘big data’analytics in the public sector”. In:
Public Money & Management36.5 (2016), pp. 385–390.
[2] Information on smart city initiatives of the City of Copenhagen.url:https://urbandevelopmentcph.
kk.dk/indhold/smart-city(visited on 02/16/2020).
[3] Open Datasets of the City of Copenhagen on Open Data DK.url:https://www.opendata.
dk/city-of-copenhagen(visited on 02/18/2020).
[4] Website of Copenhagen Solutions Lab.url:https://www.cphsolutionslab.dk/en(visited on 02/18/2020).
[5] City of Copenhagen Smart Solutions Overview. url: https : / / stateofgreen . com / en / partners/city-of-copenhagen/(visited on 06/11/2020).
[6] H. Furuta, T. Kameda, Y. Fukuda, and D. M. Frangopol. “Life-cycle cost analysis for infrastructure systems: life-cycle cost vs. safety level vs. service life”. In: Life-cycle performance of deteriorating structures: Assessment, design and management. 2004, pp. 19–25.
[7] F. K. Rioja. “Filling potholes: macroeconomic effects of maintenance versus new invest-ments in public infrastructure”. In:Journal of Public Economics87.9-10 (2003), pp. 2281–
2304.
[8] "So Many Potholes, So Much Cost" - Article from pothole.info.url:https://www.pothole.
info/2016/05/so-many-potholes-so-much-cost/(visited on 05/25/2020).
[9] Claim for Compensation after Accident on Public Road in Copenhagen.url:https://www.kk.
dk/indhold/uheld-paa-offentlig-vej(visited on 06/09/2020).
[10] M. Gkemou, A. Spiliotis, and E. Bekiaris. “Needs, restrictions and priorities for the development and installation of a multifunctional C-ITS solution on existing road pavement”. In:Transportation research procedia36 (2018), pp. 185–192.
[11] E. Brynjolfsson and A. Mcafee. “The business of artificial intelligence”. In: Harvard Business Review(2017), pp. 1–20.
[12] M. Batty, K. W. Axhausen, F. Giannotti, A. Pozdnoukhov, A. Bazzani, M. Wachowicz, G. Ouzounis, and Y. Portugali. “Smart cities of the future”. In:The European Physical Journal Special Topics214.1 (2012), pp. 481–518.
[13] A. Zanella, N. Bui, A. Castellani, L. Vangelista, and M. Zorzi. “Internet of things for smart cities”. In:IEEE Internet of Things journal1.1 (2014), pp. 22–32.
[14] A. Gharaibeh, M. A. Salahuddin, S. J. Hussini, A. Khreishah, I. Khalil, M. Guizani, and A. Al-Fuqaha. “Smart cities: A survey on data management, security, and enabling technologies”. In:IEEE Communications Surveys & Tutorials19.4 (2017), pp. 2456–2501.
[15] N/A.Vejvedligehold - vedligehold af færdselsarealet. Vejdirektoratet (Danish Road
Direc-torate, 2009.url:http://www.glstrandbjerggaard.dk/images/temaer/vejvedligeholdelse/
vej_vedligeholdelse.pdf(visited on 04/07/2020).
[16] ASTM. “Standard Practice for Roads and Parking Lots Pavement Condition Index Surveys”. In: D6433, West Conshohocken, PA. 2011.
[17] T. Okuda, K. Suzuki, and N. Kohtake. “Non-parametric Prediction Interval Estimate for Uncertainty Quantification of the Prediction of Road Pavement Deterioration”. In:
2018 21st International Conference on Intelligent Transportation Systems (ITSC). IEEE. 2018, pp. 824–830.
[18] P. Marcelino, M. de Lurdes Antunes, E. Fortunato, and M. C. Gomes. “Transfer learning for pavement performance prediction”. In:International Journal of Pavement Research and Technology13.2 (2020), pp. 154–167.
[19] N. Nabipour, N. Karballaeezadeh, A. Dineva, A. Mosavi, S. Mohammadzadeh, S.
Shamshirband, et al. “Comparative Analysis of Machine Learning Models for Prediction of Remaining Service Life of Flexible Pavement”. In:Mathematics7.12 (2019), p. 1198.
[20] N. Karballaeezadeh, F. Zaremotekhases, S. Shamshirband, A. Mosavi, N. Nabipour, P. Csiba, and A. R. Várkonyi-Kóczy. “Intelligent Road Inspection with Advanced Machine Learning; Hybrid Prediction Models for Smart Mobility and Transportation Maintenance Systems”. In:Energies13.7 (2020), p. 1718.
[21] G. E. Elkins, T. Thompson, J. Groeger, B. A. Visintine, G. R. Rada, et al.Reformulated pavement remaining service life framework. Tech. rep. United States. Federal Highway Administration. Office of Infrastructure . . ., 2013.
[22] N. Karballaeezadeh, D. Mohammadzadeh S, S. Shamshirband, P. Hajikhodaverdikhan, A. Mosavi, and K.-w. Chau. “Prediction of remaining service life of pavement using an optimized support vector machine (case study of Semnan–Firuzkuh road)”. In:
Engineering Applications of Computational Fluid Mechanics13.1 (2019), pp. 188–198.
[23] M. M. Zumrawi. “Investigating causes of pavement deterioration in Khartoum state, Su-dan”. In:International Journal of Civil and Environmental Engineering9.11 (2015), pp. 1450–
1455.
[24] G. Doré, J.-M. Konrad, and M. Roy. “Deterioration model for pavements in frost conditions”. In:Transportation research record1655.1 (1999), pp. 110–117.
[25] R. K. Mobley.An introduction to predictive maintenance, Second Edition. Elsevier, 2002.
[26] S. Selcuk. “Predictive maintenance, its implementation and latest trends”. In:Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture231.9 (2017), pp. 1670–1679.
[27] S. Russell and P. Norvig.Artificial intelligence: a modern approach, Third Edition. Pearson Education, 2010.
[28] A. Géron.Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems. O’Reilly Media, 2019.
[29] T. Hastie, R. Tibshirani, and J. Friedman.The elements of statistical learning: data mining, inference, and prediction. Springer Science & Business Media, 2009.
[30] G. James, D. Witten, T. Hastie, and R. Tibshirani.An introduction to statistical learning.
Vol. 112. Springer, 2013.
[31] C. Strobl, J. Malley, and G. Tutz. “An introduction to recursive partitioning: rationale, application, and characteristics of classification and regression trees, bagging, and random forests.” In:Psychological methods14.4 (2009), p. 323.
[32] XGBoost Documentation. url: https://xgboost.readthedocs.io/en/latest/index.
html(visited on 05/30/2020).
[33] R. J. Hyndman and G. Athanasopoulos.Forecasting: principles and practice. OTexts, 2018.
[34] C. Ji, X. Zou, Y. Hu, S. Liu, L. Lyu, and X. Zheng. “XG-SF: An XGBoost classifier based on shapelet features for time series classification”. In:Procedia computer science147 (2019), pp. 24–28.
[35] J. Bergstra and Y. Bengio. “Random search for hyper-parameter optimization”. In:
Journal of machine learning research13.Feb (2012), pp. 281–305.
[36] F. Provost and T. Fawcett. “Data science and its relationship to big data and data-driven decision making”. In:Big data1.1 (2013), pp. 51–59.
[37] E. Brynjolfsson, L. M. Hitt, and H. H. Kim. “Strength in numbers: How does data-driven decisionmaking affect firm performance?” In:Available at SSRN 1819486(2011).
[38] Organization diagram of the Technical and Environmental Management department. url: https : / / www . kk . dk / sites / default / files / teknik - _og _ miljoeforvaltningens _ organisation_010818.pdf(visited on 03/09/2020).
[39] Specification about RoSy Road Asset Management System.url: https://www.sweco.dk/vi-tilbyder / infrastruktur / ?service = Drift % 5C % 20og % 5C % 20vedligeholdelse % 5C % 20af%5C%20bel%C3%A6gninger%5C%20og%5C%20assets(visited on 04/04/2020).
[40] Overview about PUMA Data Platform. url: https : / / www . gate21 . dk / wp - content / uploads / 2017 / 11 / PUMA _ Rolf - Foxby _ K % C3 % B8benhavns - Kommune . pdf (visited on 03/15/2020).
[41] Website of Giv et praj.url:https://givetpraj.kk.dk(visited on 04/04/2020).
[42] N/A. Skadeskatalog - Retningslinjer for tilstandsregistrering. Sweco Danmark A/S.url: https://www.svendborg.dk/sites/default/files/skadeskatalog.pdf (visited on 04/07/2020).
[43] N/A.Retningslinjer for tilstandsregistreringer – kørebaner og cykelstier. City of Copenhagen.
url:https://www.kk.dk/sites/default/files/uploaded-files/Tilstandsregistrering%
5C % 20 - %5C % 20Bilag % 5C % 20C % 5C % 20 - %5C % 20Retningslinjer % 5C % 20for % 5C % 20tilstandsregistering%5C%20af%5C%20k%C3%B8rebaner%5C%20og%5C%20cykelstier.
pdf(visited on 05/03/2020).
[44] M. Saunders, P. Lewis, and A. Thornhill. Research methods for business students. 8th edition. Pearson Education Limited, England, 2019.
[45] R. Wirth and J. Hipp. “CRISP-DM: Towards a standard process model for data mining”.
In:Proceedings of the 4th international conference on the practical applications of knowledge discovery and data mining. Springer-Verlag London, UK. 2000, pp. 29–39.
[46] Overview about EPSG:3857.url:https://epsg.io/3857(visited on 03/19/2020).
[47] S. Garavaglia and A. Sharma. “A smart guide to dummy variables: Four applications and a macro”. In:Proceedings of the Northeast SAS Users Group Conference. 1998, p. 43.
[48] J. v. Brocke, A. Simons, B. Niehaves, B. Niehaves, K. Reimer, R. Plattfaut, and A. Cleven.
“Reconstructing the giant: On the importance of rigour in documenting the literature search process”. In:European Conference on Information Systems (ECIS)(2009), pp. 2206–
2217.
[49] J. Heer, M. Bostock, and V. Ogievetsky. “A tour through the visualization zoo”. In:
Communications of the ACM53.6 (2010), pp. 59–67.
[50] A. Alfarrarjeh, D. Trivedi, S. H. Kim, and C. Shahabi. “A deep learning approach for road damage detection from smartphone images”. In:2018 IEEE International Conference on Big Data (Big Data). IEEE. 2018, pp. 5201–5204.
[51] Y. J. Wang, M. Ding, S. Kan, S. Zhang, and C. Lu. “Deep Proposal and Detection Networks for Road Damage Detection and Classification”. In:2018 IEEE International Conference on Big Data (Big Data). IEEE. 2018, pp. 5224–5227.
[52] Y. Zou, W. Cao, M. Luo, P. Zhang, W. Wang, and W. Huang. “Deep Learning-based Pavement Cracks Detection via Wireless Visible Light Camera-based Network”. In:2019 Computing, Communications and IoT Applications (ComComAp). IEEE. 2019, pp. 47–52.
[53] W. Xia. “An approach for extracting road pavement disease from HD camera videos by deep convolutional networks”. In:2018 International Conference on Audio, Language and Image Processing (ICALIP). IEEE. 2018, pp. 418–422.
[54] D. Xu, G. Xu, and G. Xu. “Crack Identification Algorithm Based on MASK Dodging Principle and Deep Learning”. In:2019 Chinese Automation Congress (CAC). IEEE. 2019, pp. 5770–5776.
[55] W. Kaddah, M. Elbouz, Y. Ouerhani, V. Baltazart, M. Desthieux, and A. Alfalou. “Opti-mized minimal path selection (OMPS) method for automatic and unsupervised crack segmentation within two-dimensional pavement images”. In:The Visual Computer35.9 (2019), pp. 1293–1309.
[56] L. Zhang, F. Yang, Y. D. Zhang, and Y. J. Zhu. “Road crack detection using deep convolutional neural network”. In:2016 IEEE international conference on image processing (ICIP). IEEE. 2016, pp. 3708–3712.
[57] X. Wang and Z. Hu. “Grid-based pavement crack analysis using deep learning”. In:
2017 4th International Conference on Transportation Information and Safety (ICTIS). IEEE.
2017, pp. 917–924.
[58] N.-D. Hoang and Q.-L. Nguyen. “Automatic recognition of asphalt pavement cracks based on image processing and machine learning approaches: a comparative study on classifier performance”. In:Mathematical Problems in Engineering 2018 (2018).
[59] N.-D. Hoang and Q.-L. Nguyen. “Fast Local Laplacian-Based Steerable and Sobel Filters Integrated with Adaptive Boosting Classification Tree for Automatic Recognition of Asphalt Pavement Cracks”. In:Advances in Civil Engineering2018 (2018).
[60] A. Banharnsakun. “Hybrid ABC-ANN for pavement surface distress detection and classification”. In: International Journal of Machine Learning and Cybernetics 8.2 (2017), pp. 699–710.
[61] K. Doycheva, C. Koch, and M. König. “Implementing textural features on GPUs for improved real-time pavement distress detection”. In:Journal of Real-Time Image Processing 16.5 (2019), pp. 1383–1394.
[62] S. Wu, J. Fang, X. Zheng, and X. Li. “Sample and Structure-Guided Network for Road Crack Detection”. In:IEEE Access7 (2019), pp. 130032–130043.
[63] D. B. A. Chacra and J. S. Zelek. “Fully automated road defect detection using street view images”. In:2017 14th Conference on Computer and Robot Vision (CRV). IEEE. 2017, pp. 353–360.
[64] S. Chatterjee, P. Saeedfar, S. Tofangchi, and L. Kolbe. “Intelligent Road Maintenance: a Machine Learning Approach for surface Defect Detection.” In:ECIS. 2018, p. 194.
[65] Y. Pan, X. Zhang, G. Cervone, and L. Yang. “Detection of asphalt pavement potholes and cracks based on the unmanned aerial vehicle multispectral imagery”. In: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 11.10 (2018), pp. 3701–3712.
[66] R.-Y. Wang, Y.-T. Chuang, and C.-W. Yi. “A crowdsourcing-based road anomaly classifi-cation system”. In:2016 18th Asia-Pacific Network Operations and Management Symposium (APNOMS). IEEE. 2016, pp. 1–4.
[67] A. Basavaraju, J. Du, F. Zhou, and J. Ji. “A Machine Learning Approach to Road Surface Anomaly Assessment using Smartphone Sensors”. In:IEEE Sensors Journal20.5 (2019), pp. 2635–2647.
[68] Z. Zheng, M. Zhou, Y. Chen, M. Huo, and D. Chen. “Enabling real-time road anomaly detection via mobile edge computing”. In: International Journal of Distributed Sensor Networks15.11 (2019), p. 1550147719891319.
[69] K. Laubis, V. Simko, and C. Weinhardt. “Weighted aggregation in the domain of crowd-based road condition monitoring”. In:Informatik 2016(2016).
[70] F. Seraj, K. Zhang, O. Turkes, N. Meratnia, and P. J. Havinga. “A smartphone based method to enhance road pavement anomaly detection by analyzing the driver behavior”.
In:Adjunct Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2015 ACM International Symposium on Wearable Computers. 2015, pp. 1169–1177.
[71] M. R. Dey, U. Satapathy, P. Bhanse, B. K. Mohanta, and D. Jena. “MagTrack: Detect-ing Road Surface Condition usDetect-ing Smartphone Sensors and Machine LearnDetect-ing”. In:
TENCON 2019-2019 IEEE Region 10 Conference (TENCON). IEEE. 2019, pp. 2485–2489.
[72] T. Wickramarathne, V. Garg, and P. Bauer. “On the use of 3-d accelerometers for road quality assessment”. In:2018 IEEE 87th Vehicular Technology Conference (VTC Spring).
IEEE. 2018, pp. 1–5.
[73] H. Tariq, S. Mazhar, and H. Hameed. “Road quality classification for road repair authorities and regular drivers, using an on-board data logger”. In:2018 17th ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN). IEEE. 2018, pp. 142–143.
[74] N. Hassan, I. Siddiqui, S. Mazhar, and H. Hameed. “Road Anomaly Classification for Low-Cost Road Maintenance and Route Quality Maps”. In: 2019 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops).
IEEE. 2019, pp. 645–650.
[75] C. Guo, K. Yu, Y. Gong, and J. Yi. “Optimal motion planning and control of a crack filling robot for civil infrastructure automation”. In:2017 13th IEEE Conference on Automation Science and Engineering (CASE). IEEE. 2017, pp. 1463–1468.
[76] Dutch company uses smart technology to make roads safer. url: https : / / customers . microsoft.com/en-us/story/724143-royal-bam-azure-machine-learning (visited on 05/11/2020).
[77] Drones and Artificial Intelligence deliver substantial savings on Storebælt. url: https://
sundogbaelt . dk / en / in focus / drones and artificial intelligence deliver -substantial-savings-on-storebaelt/(visited on 05/11/2020).
[78] IBM to Develop an AI-Powered IoT Solution to Help Clients Manage and Monitor Aging Bridges, Tunnels, Highways and Railways.url: https://newsroom.ibm.com/2019- 04- 24-IBM-to-Develop-an-AI-Powered-IoT-Solution-to-Help-Clients-Manage-and-Monitor-Aging-Bridges-Tunnels-Highways-and-Railways(visited on 05/12/2020).
[79] AI in road maintenance to be deployed in UK.url: https://www.governmenteuropa.eu/ai-in-road-maintenance/94040/(visited on 05/13/2020).
[80] Blackpool Council Case Study. url: https : / / www . gaist . co . uk / case - studies / blackpool-council(visited on 05/13/2020).
[81] Vaisala - Road Condition Analysis with RoadAI. url: https : / / www . vaisala . com / en / products / data subscriptions and reports / computer vision and data -visualization/roadai-road-condition-analysis(visited on 05/13/2020).
[82] Dash-cams to be used to identify potholes and speed up repairs. url: https://www.eadt.
co.uk/news/dashcams-used-to-spot-potholes-in-suffolk-1-6171650 (visited on 05/13/2020).
[83] T. Lundberg, P. Andren, T. Wahlmann, O. Eriksson, L. Sjögren, and P. Ekdahl. New technology for road surface measurement - Transverse profile and rut depth. Swedish National Road and Transport Research Institute (VTI). url: http://vti.diva- portal.org/
smash/get/diva2:1192178/FULLTEXT01.pdf(visited on 05/13/2020).
[84] Description of Pavement Profile Scanner PPS / PPS-PLUS. url: https : / / www . ipm . fraunhofer.de/content/dam/ipm/en/PDFs/product-information/OF/MTS/Pavement-Profile-Scanner-en.pdf(visited on 05/13/2020).
[85] Presentation about Pavement Profile Scanner Plus - Next Generation in Road Monitoring, Modelling and Mapping.url: https://www.eadt.co.uk/news/dashcams-used-to-spot-potholes-in-suffolk-1-6171650(visited on 05/13/2020).
[86] Overview about Pavemetrics Laser Crack Measurement System (LCMS-2). url: http : / / www . pavemetrics . com / applications / road - inspection / lcms2 - en/ (visited on 05/13/2020).
[87] Y.-C. J. Tsai, A. Chatterjee, and C. Jiang. “Challenges and lessons from the successful implementation of automated road condition surveys on a large highway system”. In:
2017 25th European Signal Processing Conference (EUSIPCO). IEEE. 2017, pp. 2031–2035.
[88] Website of MacRebur Ltd.url:https://www.macrebur.com/(visited on 05/14/2020).
[89] S. Gleave, R. Frisoni, F. Dionori, L. Casullo, C. Vollath, L. Devenish, F. Spano, T. Sawicki, S. Carl, R. Lidia, et al. “EU road surfaces: Economic and safety impact of the lack of regular road maintenance”. In: European Parliament-Directorate General for Internal Policies, Policy Department B: Structural and Cohesion Policies, Transport and Tourism(2014).
A.1. Images of Road Damage Types
Down below are presented exemplary images of each road damage type that was described in 3.1.3. All images are obtained from [43] if not stated otherwise.
Small cracks
Large cracks
Alligator cracks
Settlements
Image obtained from [42].
Rutting
Image obtained from [42].
Raveling
Image obtained from [42].
Chip loss
Image obtained from [42].
Stripping/ Peeling
Potholes
Images obtained from [42].
Emergency reparations
Image obtained from [42].
Patches
A.2. Raw Data Examples
Down below are exemplary extracts of each data source that was used in the thesis.
Current Condition (Part 1/2)
Current Condition (Part 2/2)
History Damages (Part 1/2)
History Damages (Part 2/2)
Traffic Data
Pavement Data
RoSy Longlist
Product Lifetime and Prices (Part 1/2)
Product Lifetime and Prices (Part 2/2)
PUMA Dataset
PUMA tasks and Number of Records