MODELING SKID RESISTANCE FOR FLEXIBLE PAVEMENTS: A COMPARISON BETWEEN REGRESSION AND NEURAL NETWORK MODELS

The phenomenon of pavement surface friction or skid resistance involves the complex interaction of pavement, vehicle, and environmental factors. Results of skid resistance measurements form the basis of many pavement management safety decisions including the identification of areas of excessive slipperiness, planning of maintenance or rehabilitation activities, and evaluation of material types and new construction practices. Various model forms of skid resistance have been developed using classical statistical methods such as regression. Computational models designed to resemble the human brain and charactreized as neural networks have been used successfully in the past in other fields such as economics, medicine, and stock market research to analyze problems involving very complex interrelationships; they are found to perform better than classical statistical methods. The use of neural network models as an alternative to regression models for predicting skid resistance on flexible pavements for assessing future rehabilitation needs is examined. Using data from in-service flexible pavements, separate skid resistance models are developed with both regression and neural network methods. The models are tested and compared, and the results indicate that neural networks can model their environment more convincingly than regression models for the flexible pavements studied.

Language

  • English

Media Info

  • Features: Figures; References; Tables;
  • Pagination: p. 60-71
  • Monograph Title: Pavement-vehicle interaction and traffic monitoring
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00715516
  • Record Type: Publication
  • Files: TRIS, TRB, ATRI
  • Created Date: Dec 28 1996 12:00AM