AI-RWAY

AI-RWay: Railway Inspection Goes Intelligent

The railway sector is experiencing a growing need to automate infrastructure inspection and monitoring activities, reducing manual interventions and making processes more traceable, efficient, and repeatable.

From this need comes AI-RWay, a platform developed by ZIRAK and TEKFER for the automated inspection of the railway network. The solution uses video footage acquired by drones and georeferenced data to generate operational information and timely alerts in support of maintenance activities.

A Platform for Predictive Maintenance

AI-RWay integrates the entire inspection process: from data acquisition through to the operational management of interventions. The platform combines Artificial Intelligence, Computer Vision, and Machine Learning technologies to automatically analyse video content and detect anomalies along the railway line.

Key features include:

• automatic acquisition and management of drone video footage
• intelligent data analysis through AI models
• generation of geolocated events and alerts
• map-based visualisation and intervention management via web application

This approach enables a more advanced maintenance model, based on objective data updated in real time.

Main Areas of Application

The platform is designed to support several use cases, including:

• railway signage monitoring, with detection of missing or damaged signals
• identification of obstacles and anomalous objects on the line
• track circuit (TC) status monitoring, with identification of any critical issues

Results and Impact

The project led to the development and validation of a complete solution tested in real-world scenarios, achieving high performance:

• 94% accuracy in object and obstacle detection
• 90% in signage classification
• up to 99% in track circuit monitoring

AI-RWay thus enables the transformation of railway inspection from a manual activity into a digital, automated process — improving efficiency and supporting a structured approach to predictive maintenance. The project was carried out under the regional F.E.S.R. 2021/2027 programme, dedicated to supporting innovation and industrial research.