Using ANN modeling for pavement layer moduli backcalculation as a function of traffic speed deflections

Backcalculating pavement layer moduli is one of the traditional practices used for evaluating the structural capacity of pavements. Currently, this process is mostly based on stationary deflection measurements using the falling weight deflectometer (FWD) device that has an inherent challenge of needing traffic lane closures during testing. This paper presents a new methodology for backcalculating pavement layer moduli as a function of traffic speed deflections using Artificial Neural Network (ANN). From the study results and corresponding findings, the developed ANN model exhibited acceptable accuracy with an overall coefficient of determination (R2) of 94.08% between the targeted and predicted moduli values. In addition, the ANN backcalculated moduli were satisfactorily validated against both laboratory and field data - yielding an average root mean square error (RMSE) of 5.42% for the asphalt layer moduli, 4.33% for the base layer moduli, and 6.50% for the subgrade layer moduli.