AI-Driven Multisensor Fusion for Sustainable Autonomous Railway Infrastructure Monitoring

Purpose and goal

The goal of this project is to develop an AI-driven multi-sensor fusion system for condition monitoring of railway infrastructure. By combining data from high-resolution visual imaging and robust magnetic sensors, the solution will enhance accuracy, resilience, and reliability of track inspections. The purpose is to enable proactive, cost-efficient maintenance that extends infrastructure lifespan and supports a safe, sustainable transport system.

Expected effects and result

The project aims to deliver a validated TRL 7 prototype system, capable of continuous monitoring of railway infrastructure even in extreme Nordic conditions. The system will detect faults earlier, reduce false alarms, and improve maintenance efficiency, helping to lower costs and minimize disruptions. Broader effects include reducing carbon emissions by avoiding unnecessary repairs, extending asset life, and strengthening Sweden’s leadership in digital infrastructure and applied AI.

Planned approach and implementation

The project will install and calibrate magnetic and visual sensors on a measurement train, collecting synchronized data from both test and real tracks. Data will be quality-checked, stored, and used to develop advanced AI fusion models to for defect detection. Validation will take place through test campaigns on Swedish tracks. Close collaboration between Trafikverket, industry partners, and academia ensures that the solution is practical, scalable, and ready for future deployment.