Transfer learning for pavement performance prediction

  • Pedro MarcelinoEmail author
  • Maria de Lurdes Antunes
  • Eduardo Fortunato
  • Marta Castilho Gomes


Accurate pavement performance prediction models are essential to ensure optimal allocation of resources in maintenance management. These models are developed using inventory and monitoring data regarding pavement structure, climate, traffic, and condition.

However, numerous road agencies have limited pavement data. Due to the inexistence of historical data, data collection frequency, and/or quality issues, the amount of data available for the development of performance models is reduced. As a result, the resource allocati on process is significantly undermined.

This paper proposes a transfer learning approach to develop pavement performance prediction models in limited data contexts. The proposed transfer learning approach is based on a boosting algorithm. In particular, a modified version of the popular TrAdaBoost learning algorithm was used.

To test the proposed transfer learning approach, a case study was developed using data from the Long-Term Pavement Performance (LTPP) database and from the Portuguese road administration database.

The results of this work show that it is possible to develop accurate performance prediction models in limited data contexts when a transfer learning approach is applied. All the models resulting from this approach outperformed baseline models, especially in what regards long-term forecasts. The results also showed that the transfer learning models perform consistently over different time frames, with minor performance losses from one-step to multi-step forecasts.

The findings of this study should be of interest to road agencies facing limited data contexts and aiming to develop accurate prediction models that can improve their pavement management practice.


Transfer learning Pavement performance models International Roughness Index (IRI) Machine learning Pavement Management Systems (PMS) Predictive maintenance 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.



This work was supported by Fundação para a Ciência e a Tecnologia (FCT) [grant number SFRH/BD/129907/2017].


  1. [1]
    R. Haas, R. Hudson, L.C. Falls, Pavement asset management, John Wiley & Sons, 2015. Google Scholar
  2. [2]
    S. Gowda, S. Jain, K. S. Patil, K. P. Biligiri, B. E. Prabhakar, Development of a Budget-Prioritized Pavement Management System based on Engineering Criteria, Int. J. Pavement Res. Technol. 8 (3) (2015) 206–220. Google Scholar
  3. [3]
    P. Marcelino, M. L. Antunes, E. Fortunato, Comprehensive performance indicators for road pavement condition assessment, Struct. Infrastruct. Eng. 14 (11) (2018) 1–13. Google Scholar
  4. [4]
    S. Cafiso, A. Di Graziano, D. G. Goulias, C. D'Agostino, Distress and profile data analysis for condition assessment in pavement management systems, Int. J. Pavement Res. Technol. 12 (5) (2019) 527–536. CrossRefGoogle Scholar
  5. [5]
    T.A. Dawson, Evaluation of pavement management data and analysis of treatment effectiveness using multi-level treatment transition matrices, Michigan State University, East Lansing, Michigan, USA, 2012.Google Scholar
  6. [6]
    G. Y. Baladi, S. W. Haider, K. Chatti, L. Galehouse, T. A. Dawson, C. M. Dean, R. Muscott, N. Tecca, M. McClosky, Optimization of and Maximizing the Benefits from Pavement Management Data Collection, Federal Highway Administration, Washington, D.C, USA, 2009.Google Scholar
  7. [7]
    L.M. Pierce, G. McGovern, K.A. Zimmerman, Practical guide for quality management of pavement condition data collection, US Department of Transportation, Federal Highway Administration, Washington, D.C, USA, 2013.Google Scholar
  8. [8]
    R.L. Schmitt, S. Owusu-Ababio, K.D. Denn, Database Development for an HMA Pavement Performance Analysis System.No. MRUTC 07-11. Midwest Regional University Transportation Center, College of Engineering, Department of Civil and Environmental Engineering, University of Wisconsin, Madison, Wisconsin, USA, 2008.Google Scholar
  9. [9]
    A. Ferreira, L. Picado-Santos, Z. Wu, G. Flintsch, Selection of pavement performance models for use in the Portuguese PMS, Int. J. Pavement Eng. 12 (1) (2011) 87–97. CrossRefGoogle Scholar
  10. [10]
    L.N. Mills, N.O. Attoh-Okine, S. McNeil. Developing pavement performance models for Delaware, Transport. Res. Rec. 2304 (1) (2012) 97–103. CrossRefGoogle Scholar
  11. [11]
    A. Ayed, Development of Empirical and Mechanistic Empirical performance Models at Project and Network Levels, (PhD Thesis), University of Waterloo, Canada, 2016.Google Scholar
  12. [12]
    P. Marcelino, M. L. Antunes, E. Fortunato, M. G. Gomes, Machine learning approach for pavement performance prediction, Int. J. Pavement Eng. (2019) 1–14. Google Scholar
  13. [13]
    S.J. Pan, Q. Yang, A survey on transfer learning. IEEE Trans. Knowl. Data Eng. 22 (10) (2009) 1345–1359. CrossRefGoogle Scholar
  14. [14]
    Q. Zou, Y. Cao, Q. Li, Q. Mao, S. Wang, CrackTree: Automatic crack detection from pavement images, Pattern Recogn. Lett. 33 (2) (2012) 227–238. CrossRefGoogle Scholar
  15. [15]
    H.C. Shin, H.R. Roth, M. Gao, L. Lu, Z. Xu, I. Nogues, J. Yao, D. Mollura, R.M. Summers. Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning, In IEEE Trans. Med. Imaging. 35 (5) (2016) 1285–1298. CrossRefGoogle Scholar
  16. [16]
    K. Gopalakrishnan, S.K. Khaitan, A. Choudhary, A. Agrawal. Deep Convolutional Neural Networks with transfer learning for computer vision-based data-driven pavement distress detection, Construction and Building Materials 157 (2017) 322–330. CrossRefGoogle Scholar
  17. [17]
    K. Zhang, H.D. Cheng, B. Zhang B. Unified approach to pavement crack and sealed crack detection using preclassification based on transfer learning, J. Comput. Civil Eng. 32 (2) (2018).
  18. [18]
    B. Kim, S. Cho, Automated vision-based detection of cracks on concrete surfaces using a deep learning technique, Sensors 18 (10) (2018) 3452. CrossRefGoogle Scholar
  19. [19]
    W. R. Silva, D.S. Lucena DS, Concrete cracks detection based on deep learning image classification, In Multidisciplinary Digital Publishing Institute Proceedings, 2018. Google Scholar
  20. [20]
    K. Gopalakrishnan, H. Gholami, A. Vidyadharan, A. Choudhary, A. Agrawal, Crack damage detection in unmanned aerial vehicle images of civil infrastructure using pre-trained deep learning model, Int. J. Traffic Transp. Eng. 8 (1) (2018) 1–4. CrossRefGoogle Scholar
  21. [21]
    J. Garcke, T. Vanck. Importance weighted inductive transfer learning for regression. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, Nancy, France, 2014, pp. 466–481. Google Scholar
  22. [22]
    R.E. Schapire, The boosting approach to machine learning: An overview. In: Denison D.D., Hansen M.H., Holmes C.C., Mallick B., Yu B. (eds) Nonlinear estimation and classification. Springer, New York, USA, 2003, pp. 149–171.CrossRefGoogle Scholar
  23. [23]
    W. Dai, Q. Yang, G.R. Xue, Y. Yu, Boosting for transfer learning, In Proceedings of the 24th international conference on Machine learning, Oregon, USA, 2007, pp. 193–200. Google Scholar
  24. [24]
    H. Drucker, Improving regressors using boosting techniques, In Proceedings of the 14th International Conference on Machine Learning, Nashville, USA, 1997, pp. 107–115.Google Scholar
  25. [25]
    D. Pardoe, P. Stone, Boosting for regression transfer, In Proceedings of the 27th International Conference on Machine Learning, Haifa, Israel, 2010, pp. 863–870.Google Scholar
  26. [26]
    Y. Freund, R. Schapire, A decision-theoretic generalization of online learning and an application to boosting. J. Comput. Syst. Sci. 55 (1997) 119–139. zbMATHCrossRefGoogle Scholar
  27. [27]
    A. Blum, T. Mitchell, Combining labeled and unlabeled data with co-training, In Proceedings of the 11th Annual Conference on Computational Learning Theory, Madison, USA, 1998, pp. 92–100.Google Scholar
  28. [28]
    M. Belkin, P. Niyogi, V. Sindhwani, Manifold regularization: A geometric framework for learning from labeled and unlabeled examples. J. Mach. Learn. Res. 7 (Nov) (2006) 2399–434.Google Scholar
  29. [29]
    X. Zhu, Z. Ghahramani, Learning from labeled and unlabeled data with label propagation, Carnegie Mellon University, Pittsburgh, USA, 2002.Google Scholar
  30. [30]
    M.W. Sayers, S.M. Karamihas, The Little Book of Profiling, University of Michigan, Transportation Research Institute, Ann Arbor, Michigan, USA, 1998.Google Scholar
  31. [31]
    K. Golabi, P. Pereira, Innovative pavement management and planning system for road network of Portugal. J. Infrastructure Syst. 9 (2) (2003) 75–80. CrossRefGoogle Scholar
  32. [32]
    L. Picado-Santos, A. Ferreira, Development and implementation of a new pavement management system. In Proceedings of the 5th International Symposium on Maintenance and Rehabilitation of Pavements and Technological Control, Iowa City, USA, 2007, pp. 433–438.Google Scholar
  33. [33]
    L. Picado-Santos, A. Ferreira, Contributions to the development of the Portuguese road administration's pavement management system, In Proceedings of the 3rd European Pavement and Asset Management Conference 1138, Coimbra, Portugal, 2008, pp. 1–10.Google Scholar
  34. [34]
    A. Luz, L.G. Picado-Santos, Contribution to the quality index model in the Description of the national road network, Universidade de Lisboa, Instituto Superior Técnico, Lisbon, Portugal, 2011.Google Scholar
  35. [35]
    P. Marcelino, M. L. Antunes, E. Fortunato, Current international practices on pavement condition assessment, In Pavement and Asset Management: Proceedings of the World Conference on Pavement and Asset Management (WCPAM 2017), Milan, Italy, 2017.Google Scholar
  36. [36]
    S. Nassiri, M. H. Shafiee, A. Bayat, Development of roughness prediction models using Alberta transportation's pavement management system, Int. J. Pavement Res. Technol. 6 (6) (2013) 714–720. Google Scholar
  37. [37]
    Federal Highway Administration, National Performance Management Measures, Federal Register 80 (2), Washington, D.C, USA, 2015, pp. 326–393.Google Scholar
  38. [38]
    LTPP DataPave online, Long-Term Pavement Performance, (LTPP, 2018). Database. Accessed May 2018.Google Scholar
  39. [39]
    J. Osborne, Notes on the use of data transformations. Pract. Assess. Res. Eval. 8(1) (2002) 42–50.Google Scholar
  40. [40]
    L. Breiman, Random forests. Mach. Learn. 45 (1) (2001) 5–32. zbMATHCrossRefGoogle Scholar
  41. [41]
    R.E. Schapire, The strength of weak learnability, Mach. Learn. 5 (2) (1990) 197–227. Google Scholar
  42. [42]
    F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, Scikit-learn: Machine learning in Python, J. Mach. Learn. Res. 12 (9) (2011) 2825–2830.MathSciNetzbMATHGoogle Scholar
  43. [43]
    R.W. Perera, S.D. Kohn, Issues in pavement smoothness, National Research Council, Transportation Research Board, Washington, D. C., USA, 2002.Google Scholar
  44. [44]
    N. Abdelaziz, R. El-Hakim, S. El-Badawy, H. Afify, International Roughness Index prediction model for flexible pavements. Int. J. Pavement Eng. 21 (1) (2018) 1–12. Google Scholar
  45. [45]
    G. James, D. Witten, T. Hastie, R. Tibshirani, An introduction to statistical learning, Springer, New York, USA, 2013.[dataset]zbMATHCrossRefGoogle Scholar

Copyright information

© Higher Education Press Limited Company 2019

Authors and Affiliations

  • Pedro Marcelino
    • 1
    Email author
  • Maria de Lurdes Antunes
    • 1
  • Eduardo Fortunato
    • 1
  • Marta Castilho Gomes
    • 2
  1. 1.National Laboratory for Civil EngineeringLisbonPortugal
  2. 2.CERIS, Instituto Superior TécnicoUniversidade de LisboaLisbonPortugal

Personalised recommendations