LTG Tests 3D Scanning and AI Obstacle Detection in Lithuania

VILNIUS, Lithuania – LTG Group, the state railway operator, has begun testing a package of sensor-based, AI-driven technologies that could later combine into an ecosystem for autonomous trains. The group’s first projects involve 3D scanning of the entire Lithuanian rail network—over 3,500 km of track—and artificial intelligence systems designed to detect trees, vehicles, or other obstructions on the tracks in real time. No timeline for full driverless operation has been set.

What Are the Technical Specifications?

The programme currently separates five distinct functions: automatic door-area monitoring, real-time route surveillance and obstacle detection, speed and traffic-parameter control, remote operational monitoring, and automated infrastructure inspections. The 3D scanning system captures the network’s geometry to build a high-fidelity digital twin, allowing LTG to identify millimetric changes in track condition, equipment, and structures. The obstacle-detection AI is being trained to analyse specific line segments continuously, flagging unexpected objects and triggering alerts to dispatchers or field staff without human observation.

Door-monitoring technology draws on cameras, sensors, and analysis algorithms to determine when boarding is complete and whether people or objects remain in the danger zone. That data can be passed automatically to the train’s departure-authorisation system. LTG’s Chief Technology Officer Vytautas Bitinas emphasised that each component will undergo isolated safety and reliability testing, first in controlled environments and later under real operating conditions, including varied weather, before any integration.

Key Technical Data

Where Does This Technology Stand in the Market?

The global market for railway signalling and sensor-based monitoring systems is projected to grow at a compound annual rate of approximately 6.8% from 2026 to 2035, reaching a market index of 185 (2025 base = 100), according to a March 2026 forecast from IndexBox. That expansion is driven partly by edge-enabled SCADA and web-based DIAView deployments that extend beyond traditional signalling into distributed energy and water-infrastructure monitoring—creating a parallel ecosystem in which LTG’s combined digital-twin and obstacle-detection programme can sit.

In terms of established autonomous-train systems, LTG’s approach differs sharply from those already in revenue service. Alstom’s ATO over ETCS has been demonstrated at GoA2 on mainline freight and passenger operations in France and the Netherlands, with automatic acceleration, braking, and door control, while a driver remains present. Siemens Mobility’s ATO over ETCS has operated on the Thameslink core in London since 2019, achieving 24 trains per hour through automation of speed and braking profiles. LTG is not yet testing train-control automation; it is instead building the perception and infrastructure-condition layers first, a sequence that Bitinas acknowledged will face lengthy regulatory approval before any functions are transferred from driver to machine. A comparable early-stage project is SNCF’s autonomous freight train programme, which began with sensor suites and image recognition on dedicated test tracks, but LTG’s explicit segmentation into five isolated functional blocks is uncommon among European state railways.

Editor’s Analysis

LTG’s stepwise separation of autonomous functions—rather than pursuing a single integrated platform—reduces the risk profile in a sector where safety-certification timelines often exceed a decade. It also aligns the operator with the broader infrastructure-investment pattern seen in the wider signalling market, where retrofitting existing lines with digital-condition monitoring is generating faster returns than greenfield automation. Lithuania’s 3,500 km network is relatively compact, making a full digital twin achievable with modest sensor density, but the programme’s ultimate success will depend on whether the individual building blocks can be proven independently reliable enough to meet the EU’s TSIs under the Fourth Railway Package.

FAQ

Q: When will fully autonomous trains run in Lithuania?
A: No date has been set. LTG describes driverless trains as a distant prospect and is currently testing separate component technologies that will need years of safety approval before any partial automation enters service.

Q: What exactly is LTG’s digital twin project?
A: It uses 3D scanning to record the geometry of the entire 3,500 km rail network, creating a continuously updateable virtual model that allows detection of even minor track or equipment changes for predictive maintenance.

Q: How does the AI obstacle-detection system work?
A: AI-equipped cameras and sensors monitor designated track sections in real time. When a fallen tree, vehicle, or other unexpected object appears, the system classifies it and sends an alert to dispatchers or field staff, cutting the time between hazard appearance and response.

Q: Who is supplying the technology for LTG’s autonomous train programme?
A: LTG has not publicly identified its technology suppliers. The group runs the tests internally and with undisclosed partners.