Machine Learning based On-Site Testing and Inspection of Asphalt Pavement Construction in Highway Engineering

Asphalt pavement quality is critical in maintaining highway safety, longevity, and performance and thus on-site inspection and testing is a critical aspect of highway engineering. Conventional inspections that are usually manual and subjective consume time, are subject to human error and inconsistency, resulting in poor construction quality and higher maintenance expenses. The current automated systems also have limitations in responding to different site conditions and making correct real-time assessments. To overcome these challenges, this research suggests a data-driven on-site testing and inspection model with the Random Forest algorithm. The suggested model combines sensor-based and visual data to detect pavement defects and analyze construction quality efficiently. Due to its high robustness and capacity to deal with complicated, nonlinear data, Random Forest is utilized to enhance prediction accuracy and decision-making at the construction stage. A thorough performance analysis was performed with the help of the important parameters like accuracy, precision, recall, specificity, and F1-score. The findings show that the Random Forest model exhibits excellent evaluation performance with 96% accuracy, 96.01% precision, 96% recall, 96.1% specificity, and 96% F1-score. These findings show that the model has good efficacy in allowing accurate, consistent, and efficient on-site testing of asphalt pavement in highway engineering.