LTTS Deploys AI-enabled Compute Vision for Rail Track Defect Detection

Background and Client 

Rail networks are experiencing higher traffic, with more frequent trains operating at greater speeds and carrying heavier loads. As a result, tracks are subject to increased wear, requiring more frequent inspection and maintenance, which can raise associated costs. Deteriorating track conditions may lead to delays, malfunctions, or derailments. Regular inspections are crucial to prevent derailments, which account for 60 percent of train accidents.

The client, a global rail wagon manufacturer sought a compact, battery-operated rail track defect detection system, which can be retrofitted into a train’s unpowered end wagon. L&T Technology Services (LTTS) was chosen for its extensive railroad experience and leadership in advanced rail inspection technologies.

What LTTS Enabled

LTTS designed and developed a unique, embedded AI-based, battery-operated, lightweight End-of-Train (EoT) mounted Broken Rail Detection System (BRDS) system, which provided real-time rail defect detection and reporting capable of operating up to 24 hours in a single charge. LTTS’ Compute Vision Track Eye® enabled accurate and real-time determination of the size and depth of the defect, enabling real-time strategizing of train routes based on the track condition.

Compute Vision TrackEi™ System

• The system employs deep neural networks for intelligent learning of surface defects.
• It features a closed-loop algorithm for precise depth measurement and defect mapping.
• On-board analytics distinguish and log defects for further analysis.

The LTTS Edge

• Embedded AI-based LTTS’ Compute Vision Track Eye® Broken Rail Detection System (BRDS).
• Investments in infrastructure such as the test-rig facility at Mysore, which enables end-to-end testing of the solution.

 How the Client Benefited

The client was satisfied as LTTS’ solution enabled 95 percent track defect detection.  LTTS keeps pace with evolving standards to offer customers the latest track inspection technology advances:

• Artificial intelligence and machine learning based defect detection.
• One system to detect both rail surface and track geometry.
• Proprietary tools and frameworks.