LTG Group Tests AI Obstacle Detection on 3,500 km Network

VILNIUS, Lithuania – LTG Group, the state-owned railway group, began testing three core automation technologies in July 2026: 3D infrastructure scanning with digital twin creation, real-time AI obstacle detection, and automated door area monitoring, covering the country’s entire 3,500 km rail network. The group has not disclosed the total investment, and fully autonomous trains remain a distant prospect, according to LTG’s chief technology officer Vytautas Bitinas.

What Are the Technical Specifications?

LTG has not published sensor specifications, scanning speeds, or algorithm accuracy for any of the three systems. The 3D scanning system aims to record network geometry and enable a highly accurate digital twin for predictive maintenance. The AI obstacle detection system is designed to identify fallen trees, vehicles, or other hazards in real time and send alerts to dispatchers. The door monitoring system uses cameras and sensors to assess boarding completion and detect people or objects in danger zones.

Key Technical Data

Where Does This Technology Stand in the Market?

Rail infrastructure monitoring using digital twins and AI is already operational on major European networks. The UK’s Network Rail deploys its Plain Line Pattern Recognition (PLPR) system on measurement trains, capturing 700,000 images per mile and using AI to detect defects (Source: Network Rail, 2020). Siemens Railigent offers cloud-based asset monitoring and predictive analytics for rail operators. LTG’s approach starts with stationary and train-mounted sensors rather than dedicated inspection trains, a lower-cost entry point. The global railway signalling and automation market, which underpins such sensor-based monitoring, is projected to grow at a compound annual growth rate of 6.8% from 2026 to 2035 (Source: IndexBox, 2026). Meanwhile, other railway groups are advancing on signalling overhauls: Wabtec is deploying Positive Train Control on Brazilian mixed-traffic railways to enforce speed limits and prevent collisions (Source: Wabtec, July 2026), and Network Rail Consulting secured a $26 million management contract with Sound Transit in the US (Source: RailUK, July 2026). LTG’s focus is on the foundational digital infrastructure before any train control automation.

Editor’s Analysis

LTG’s piecemeal automation — from door monitoring to infrastructure scanning — mirrors the safety-first doctrine in European rail, where each new function must be validated under strict national and EU regulations. The digital twin, if successfully built for a 3,500 km network, could become a template for other small-state railways, but without disclosed performance metrics or public cost-benefit analysis, its scalability remains uncertain. The current investment climate, with the railway signalling market index set to rise by 85% over a decade, suggests that early-adopter advantages in data acquisition may be significant.

FAQ

Q: What specific AI systems is LTG testing?
A: LTG is testing AI systems that detect trees, vehicles, and other obstacles on tracks in real time, as well as sensor-based door area monitoring algorithms. Technical details such as camera resolution and algorithm type have not been disclosed.

Q: Will Lithuania have driverless trains soon?
A: No. LTG’s chief technology officer stated that driverless trains remain a distant prospect; the current projects focus on individual automated functions like obstacle detection and remote monitoring.

Q: How does this project compare to other European railway automation initiatives?
A: LTG’s initiative is at an early testing stage, whereas larger networks like Network Rail in the UK have operational AI-based track inspection trains. The project is part of a broader trend toward digital twins and predictive maintenance, but no timeline for operational autonomous trains has been set.