| Abstrakt: |
The International Roughness Index (IRI) is the most popular index used to measure road surface roughness. In the current study, three integrated approaches based on Multiple Linear Regression (MLR), Artificial Neural Networks (ANNs), and Fuzzy Inference System (FIS) techniques were conducted to develop linear and nonlinear regression models using IRI and pavement distress parameters. The pavement distress data were collected on 19 roads in the St. John’s road network in Newfoundland, Canada, using a network “TotalPave” application. Several significant variables related to surface pavement distress were included as input parameters to develop the correlation between the IRI and pavement distress variables; eight input parameters included rutting, fatigue cracking, block cracking, longitudinal cracking, transverse cracking, potholes, patching, and delamination. The performance of the three techniques used in this study was evaluated using the coefficient of determination (R2), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). The results of the models revealed that the MLR, ANN, and FIS models could accurately predict IRI. According to ANNs, a coefficient of determination indicated that the correlation was increased by 60.7%, 46.5%, 12.34% and 11.01%. While RMSE was reduced by 73.6%, 78.7%, 51.32%, 70.4%, and MAE was reduced by 71.7%, 73.6%, 47.5%, and 61.1% compared to MLR and FIS, respectively. As a result, the ANN model indicated a better prediction of IRI for a given set of pavement distress parameters than FIS and MLR techniques. |
Full text from SpringerLink