The Pavement Condition Index Prediction Method Based on the PSO-SVM Model

With more highways built in China and more Chinese traveling on them, asphalt pavement is becoming increasingly difficult to repair and maintain. To address the pavement performance prediction problems in this essay, we will present prediction methods for PCI using SVM, BPNN, and PSO-SVM. Then, all these influencing factors are analyzed using a random forest model to determine their importance. Using field data from a portion of an ordinary highway and a section of an expressway, this investigation develops a pavement performance prediction model. It examines how factors such as road age, AADT, average annual temperature, annual precipitation, and relative humidity affect PCI values. From the results, we can see that the PSO-SVM model is accurate and stable for nonlinear, high-dimensional data and has strong generalization performance. As shown by the Random Forest, PCI is influenced by factors such as road age, traffic, and temperature. This will help maintain the roads' pavement. It provides sound equipment for highway preservation administration, enabling better use and greater longevity on roads.