A Comprehensive Literature Review and Bibliometric Analysis of the Development and Difficulties of Artificial Intelligence Methods for International Roughness Index Prediction

Abstract

The International Roughness Index (IRI) is a key indicator for pavement condition assessment and maintenance prioritization. In recent years, artificial intelligence (AI) techniques have been increasingly applied to improve IRI prediction accuracy; however, existing studies remain fragmented, limiting a comprehensive understanding of model performance, data sources, and practical implementation challenges. This study presents a systematic literature review and bibliometric analysis to synthesize recent developments in AI-based IRI prediction. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, 49 peer-reviewed articles published between 2008 and 2025 were analyzed using VOSviewer. The review addresses four main aspects: prediction model performance, data source effectiveness, implementation challenges, and research collaboration patterns. The results show that both machine learning methods (e.g., Random Forest, XGBoost, Gradient Boosting) and deep learning approaches (e.g., neural networks and Long Short-Term Memory) achieve high predictive accuracy, with R² values frequently exceeding 0.90. Machine learning models offer advantages in interpretability and computational efficiency, while deep learning models perform better with large and complex datasets. Bibliometric analysis identifies six major research clusters and a clear temporal evolution from early modeling studies to advanced AI applications. Despite technological progress, challenges remain, including data heterogeneity, limited interpretability, computational demands, and integration with pavement management systems. This study provides an evidence-based overview of current trends and identifies key research directions to support the practical and sustainable adoption of AI-driven IRI prediction in pavement construction and management. 

Received: 23 April 2025 | Revised: 15 December 2025 | Accepted: 3 March 2026

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work. 

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study. 

Author Contribution Statement

Lendra Lendra: Conceptualization, Methodology, Software, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization. Mochamad Agung Wibowo: Conceptualization, Validation, Formal analysis, Writing – review & editing, Supervision, Project administration, Funding acquisition. Jati Utomo Dwi Hatmoko: Validation, Writing – review & editing, Supervision. Rony Teguh: Validation, Writing – review & editing, Visualization.