Comparative Evaluation of Distress Prediction Modeling of Village Roads in India Using Regression and ANN Techniques

Abstract

In this paper, an attempt has been made to develop distress prediction models covering varying types of distresses—namely, cracking, potholes, and roughness— prevalent on the low-volume flexible pavements found in the village roads of India. In this regard, a total of 173 test sections were identified spread over the state of Tamil Nadu and the relevant data were collected. The aforementioned distress data were physically measured on these pavement sections. Thereafter, distress prediction models were developed using regression and artificial neural network (ANN) techniques. The developed models can be regarded as one of the effective tools for arriving at maintenance management strategies for village roads (low-volume roads) in India.