Artificial intelligence applications in porous asphalt pavement

Abstract

Porous asphalt pavement (PAP) was developed as a sustainable pavement solution to improve rainwater infiltration, reduce inundation, and improve driving safety. The performance of PAP is influenced by the complex interactions between pore structures, clogging mechanisms, and structural and thermal responses, thus encouraging the application of artificial intelligence (AI) in its modeling. This study aims to comprehensively synthesize the application of AI in PAP through the systematic literature review (SLR) approach. The SLR process was carried out against the publication of the 2021–2026 period from the Elsevier, SpringerLink, and MDPI databases, which resulted in 10 high-quality articles for analysis. The results show that AI enables accurate and non-destructive prediction of hydraulic and clogging performance, while AI-integrated mechanistic models achieve a correlation coefficient of up to 0.994 for structural response and an error below 5% for thermal conductivity. Image-based deep learning and generative models further demonstrate potential for pavement condition assessment and porous microstructure modeling. The novelty of this review lies in integrating evidence on AI applications across the hydraulic, clogging, structural, thermal, and microstructural domains of PAP. The findings highlight the need for an integrated, material-specific AI framework to support sustainable PAP design, monitoring, and maintenance.