AI-Vahan: An Infrastructure Intelligence Framework for AI-Enabled Road Digital Twins and Lifecycle Transportation Asset Management

Introduction: Transportation agencies require proactive, data-driven approaches to manage ageing road infrastructure throughout its lifecycle. Although advances in artificial intelligence (AI), multi-modal sensing, edge computing, and Digital Twins have improved infrastructure monitoring, existing studies remain fragmented across sensing technologies and decision-support systems, limiting their ability to transform heterogeneous observations into standardized engineering intelligence for transportation asset management. This study proposes AI-Vahan, a Design Science Research (DSR)-based framework for AI-enabled road infrastructure intelligence.

Methods: AI-Vahan was developed through systematic evidence synthesis, conceptual modelling, reference architecture development, and engineering formulation. The framework integrates seven functional layers encompassing multi-modal sensing (RGB camera, LiDAR, GNSS, and IMU), infrastructure perception, sensor fusion and edge intelligence, infrastructure intelligence, Road Digital Twin synchronisation, governance and decision intelligence, and stakeholder interaction. It introduces Geo-Referenced Infrastructure Observations (GRIOs) together with the Road Infrastructure Health Index (RIHI) and Maintenance Priority Index (MPI) to standardize infrastructure assessment. The framework was evaluated through conceptual, architectural, and engineering formulation validation and demonstrated through an illustrative application on a representative 1 km municipal road corridor along Keshiary Road, Kharagpur, India, rather than experimental field deployment.

Results: The proposed framework systematically transforms heterogeneous infrastructure observations into standardized engineering intelligence supporting Road Digital Twin synchronisation and governance-oriented decision-making. The illustrative application demonstrated the complete operational workflow from infrastructure observation and GRIO generation to engineering indicator extraction, infrastructure intelligence, maintenance prioritisation, and lifecycle-oriented transportation asset management, confirming the logical consistency and engineering completeness of the proposed framework.

Conclusions:AI-Vahan establishes a scientifically grounded framework linking AI-enabled multi-modal sensing, Infrastructure Intelligence, and Road Digital Twins for proactive transportation asset management. The framework provides a robust foundation for future prototype development, computational benchmarking, field validation, and integration within municipal transportation asset management systems.