Overview of the neural network based technique for monitoring of road condition via reconstructed road profiles

Abstract

A healthy road transport systems is essential for any country's social and economic development. It is generally observed that if road deterioration is allowed to increase, the economy will need significantly larger expenditures in subsequent years to keep the road maintenance backlog constant. This paper is part of a larger study whose main purpose is to investigate the dynamic behaviour of vehicles on the roads and how the observed behaviour can be effectively combined with other factors measured on the road and driver to assess the integrity of road and vehicle infrastructure. In this paper, vehicle vibration data are applied to an artificial neural network to reconstruct the corresponding road surface profiles. The results show that the technique is capable of reconstructing road profiles within an error margin of 45 percent while with careful control of principal error sources, the technique may achieve 20 percent error margin. This is considered to be reasonable enough for application to condition monitoring of unpaved roads servicing heavy vehicles