Israeli company Rail Vision has obtained a patent from the European Patent Office (EPO) for its advanced AI-based railway collision avoidance method and system, thereby strengthening its position in the European railway safety technology market.
The European patent covers an innovative solution that uses forward-looking electro-optical imaging, in single or multispectral spectrum, combined with a proprietary system architecture and advanced deep learning-based environment analysis.
According to the company, the technology enables accurate identification of the railway track ahead of the locomotive or train, followed by real-time detection of potential obstacles in its vicinity.
Real-time detection and advanced decision support
The Rail Vision solution uses a two-stage process based on convolutional neural networks (CNN). In the first phase, the system accurately determines the train’s running line, and in a second phase, another CNN module analyzes the area adjacent to the track to identify obstacles or risky situations.
The system is designed to generate critical alerts in the event of hazards such as:
- the appearance and condition of switches;
- obstacles on the line;
- end-of-rail effects;
- other objects or situations with collision potential.
With its advanced scene understanding capabilities, the technology provides decision support for locomotive engineers in staffed operations, but can also support fully automated decision-making processes in driverless trains.
Strengthening global intellectual property strategy
The granting of the European patent is part of Rail Vision’s strategy to expand and strengthen its global intellectual property protection, with the company having previously obtained similar approvals in the United States, Japan and India.
With this European recognition, Rail Vision strengthens its position as a provider of technological solutions for modernizing railway operations through the use of artificial intelligence, advanced sensors, and data analytics.
Share on:
