Roughness Modeling for Asphalt Overlay on Concrete Pavements Using Neural Networks

Pavement performance modeling is an important part of pavement management systems (PMS). The International Roughness Index (IRI) is accepted as an important indicator of pavement performance and used as the standard for the pavement roughness. The objective of this study is to develop pavement roughness models for asphalt overlay on concrete pavements using the Long-Term Performance Pavement (LTPP) database. A feed-forward Artificial Neural Networks (ANNs) approach with backpropagation learning algorithm was used in this study. A total of 592 data points from 52 pavement sections were analyzed. Five models were developed, and the best performing model was selected based on the lowest average square error (ASE), lowest mean absolute relative error (MARE), and highest coefficient of determination (R²). The best performing model utilizes 14 input variables (i.e. Initial IRIMean, Age, Wet-Freeze, Wet Non-Freeze, Dry-Freeze, Dry Non-Freeze, Asphalt Thickness, Concrete Thickness, CN Code, ESAL, Annual Air Temperature, Freeze Index, Freeze-Thaw, and Precipitation) and one output variable (IRIMean). Literature review indicated that roughness prediction models did not consider maintenance and rehabilitation (M&R) history as an independent variable. The use of Construction Number (CN) in the model development resulted in more realistic models considering that M&R actions affect the future condition of the pavement. Furthermore, the developed models predict future IRI values without using distress data. Therefore, the developed ANN roughness model allows local and state agencies to save time in data collection and processing, accordingly, reducing costs by providing a tool for better condition assessment and effective M&R scheduling.


  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; References; Tables;
  • Pagination: 19p

Subject/Index Terms

Filing Info

  • Accession Number: 01764418
  • Record Type: Publication
  • Report/Paper Numbers: TRBAM-21-04228
  • Files: TRIS, TRB, ATRI
  • Created Date: Dec 23 2020 11:05AM