Modeling of Flexible Pavement Serviceability Based on the Fuzzy Logic Theory

Abstract

Pavement serviceability plays a vital role in the transportation infrastructure system management. Inaccuracy in prediction leads to an improper maintenance strategy. Thereby, this research is an attempt to implement one of the soft computing methods in pavement serviceability evaluation. The fuzzy inference system (FIS) has been applied to Al-Samawah–Nasiriyah roadway as a case study because this roadway, as the southern part roadways of Iraq, suffers from the most common types of distresses such as rutting, bleeding, longitudinal and transverse cracking, and lane to shoulder drop off. The densities of these deteriorations for the different severity levels in each section are considered as inputs for the developed fuzzy model, whereas the outputs are the pavement condition index (PCI), analyzed by the Micro-Paver 5.2 software, other than the subjective variable of the required maintenance strategy which is based on expert opinion. The outcomes of the developed FIS model provide promising results with a 4% of error [the root mean squared error (RMSE) is found to be 4.21, and the mean absolute error (MAE) is 3.98].