Emphasizing Statistical Relationships Between Pavement Surface Roughness Index and Subgrade Ground Properties on Spatial Feature Extraction

The structural assessment of asphalt pavements is critical for efficient road maintenance and resource allocation. Traditional techniques, such as Falling Weight Deflectometer (FWD) testing, while accurate, are costly, labour-intensive, and disruptive to traffic flow, presenting major draw backs. This study explores the potential of estimating the structural capacity of asphalt pavements by leveraging functional parameters, specifically the International Roughness Index (IRI), a widely accessible and cost-effective metric for surface roughness. Utilizing data from six road sections across Ghana, totalling 219 km, the research develops a framework for predicting the Effective Structural Number (SN_eff) based on decomposed IRI data. The Robbins-Monro Kalman Filter was utilized to decompose IRI measurements into trend and random components, effectively isolating latent patterns obscured in raw data. Additionally, simple logarithmic transformations were applied to optimize data normalization for improved correlation analysis. The findings indicate that the decomposed IRI data exhibits significantly stronger correlations with SNeff compared to raw IRI. While individual road sections demonstrated varying correlation strengths, the combined dataset,which includes diverse traffic, material, and climatic conditions,yielded a more predictive relationship.

Keywords: Spectra analysis, Fallingweight deflectometer, International roughness index, Flexible pavements, Structural evaluation, Decomposition.