Indigenous Contactless ITMS for Indian Railways

Smart India Hackathon (SIH) 2025 – Hardware Edition

Turning every train into an intelligent track-monitoring train using IoT, AI/ML, and edge computing.

Problem Statement Details

Problem Statement ID: SIH25020

Problem Statement Title: Development of indigenous contactless Integrated Track Monitoring Systems (ITMS) for Track Recording on Indian Railways

Theme: Smart Automation Category: Hardware Team Name: NIRMAAN Domain: Railways, IoT, AI/ML, Embedded Systems

The Indian Railways requires accurate, real-time diagnostic tools to assess the health of its extensive rail network. Existing solutions rely heavily on a small number of expensive Track Recording Cars (TRCs), which leads to limited coverage, infrequent inspections, and delayed detection of critical track defects.

Project Overview

Our solution proposes a compact, indigenous contactless Integrated Track Monitoring System (ITMS) that can be retrofitted under any train coach. This “Mini-TRC” concept converts regular coaches into monitoring units, enabling continuous and real-time track-health monitoring without requiring dedicated TRC trains.

Compact Contactless ITMS Kit: A bolt-on hardware kit mounted under any coach, capable of multi-modal sensing, GNSS/GPS tagging, on-edge preprocessing, and hybrid connectivity.
Early Defect Detection: Multi-sensor fusion detects surface cracks, misalignment, rail wear, missing fasteners/bolts, ballast issues, and abnormal vibrations before they become critical.
Retrofit Philosophy: No special TRC rake required. Any regular train coach becomes a mobile track recorder.
Data Integration & Analytics: Collected data is sent to a centralized backend, integrated with existing Indian Railways maintenance systems, and used for historical trend analysis and predictive maintenance.
Goal: Provide real-time, comprehensive track-health data including gauge alignment, wear, cracks, fastener condition, and ride quality for both broad-gauge and narrow-gauge tracks.

Technical Approach

1. Sensor Suite (Multi-Modal Sensing)

Dual laser triangulation sensors for track alignment, twists, and rail wear profiles.
Inertial Measurement Units (IMUs) and accelerometers for vertical and horizontal unevenness and ride quality.
High-speed industrial cameras to inspect missing clips, bolts, fasteners, and ballast conditions.
LiDAR for clearance checks and maintaining safe distance from nearby structures.
Additional sensors: ultrasonic, temperature, strain, humidity, and GPS/GNSS for comprehensive monitoring.

2. Hardware & Compute Platform

Jetson Nano for on-board AI inference and computer vision tasks.
Raspberry Pi / similar SBC for data logging, networking, and device-level control.
Rugged EMI-shielded enclosure with vibration isolation mounts, cooling fans, and reliable power input from TRC auxiliary supply or local battery pack.
Modular design that allows sensors and compute units to be upgraded as technology evolves.

3. Data Acquisition & Edge Processing

Sampling at approximately 0.25 m resolution (every 25 cm), supporting train speeds from 20 km/h to 120 km/h.
On-board time synchronization using GPS and RTC for accurate geo-tagging of each measurement.
Preprocessing steps: noise filtering, signal conditioning, defect feature extraction (geometry, wear, vibration signatures, thermal anomalies).
Data fusion combining lasers, IMUs, cameras, LiDAR, and other sensors to create a unified defect map along the track.

4. Software, AI & Cloud

Computer vision and detection using YOLO, custom CNNs, and OpenCV-based preprocessing pipelines.
Ride quality assessment from accelerometer and vibration data.
Rail-wear profile extraction using LiDAR/laser scanner data.
System integration using Python and ROS for modular data flow and communication.
Dataset labelling and management using tools like Roboflow.
Cloud-based analytics stack for anomaly detection, predictive maintenance, and long-term trend analysis.

5. Prototype Setup

The prototype uses Jetson Nano and Raspberry Pi with industrial cameras, IMU, GPS, LiDAR, and supporting power electronics housed in a robust enclosure. The design is fully modular and runs predominantly on open-source software to keep the system cost-effective, adaptable, and easy to maintain.

Feasibility & Viability

Technical Feasibility

Uses low-cost, widely available sensors (ultrasonic, accelerometers, cameras, GPS, LiDAR).
Embedded platforms such as ESP32, STM32, Jetson Nano, and Raspberry Pi are capable of realtime preprocessing and AI inference.
Hybrid communication: LoRa for rural stretches and Wi-Fi/4G for urban sections.
Can operate with solar power plus battery backup for power independence.

Operational Viability

Compatible with both Broad-Gauge and Narrow-Gauge lines.
Rugged, compact, and modular design suitable for harsh Indian railway environments.
Highly scalable: from a prototype costing ~₹80k–₹100k to a large deployment fleet across multiple trains and zones.

Use Cases

Indian Railways Safety Monitoring: Detect cracks, weld failures, misalignments, and hotspots to prevent derailments.
Predictive Maintenance: Early alerts to engineers, reducing downtime and unplanned disruptions.
Urban Metro Systems: High-frequency monitoring for metro and rapid transit networks.
Private & Freight Corridors: Ensuring reliability and safety of cargo routes.
Export Potential: Low-cost kits for developing countries with large railway networks in Africa and Southeast Asia.

Challenges & Solutions

Sensor Noise & Reliability: Use digital filtering, sensor fusion, and robust calibration routines.
Ruggedization: IP65 enclosures, industrial-grade components, and vibration-damping mounts.
Connectivity Gaps: Hybrid approach with LoRa + 4G/Wi-Fi and local SD card backup to avoid data loss.
Maintenance of IoT Kits: Regular calibration, self-check algorithms, and a dedicated maintenance dashboard.
Adoption Resistance: Pilot deployments with RDSO and Indian Railways to demonstrate safety and cost benefits.

Business Potential

Indian Railways’ safety & maintenance market is estimated at ₹5,000+ Cr annually.
Potential to replace or complement TRCs using thousands of retrofit kits across the network.
B2G model: Supply kits to Indian Railways, Metro corporations, and freight operators.
Subscription-based cloud dashboard and predictive maintenance reporting services.
Export of cost-effective, indigenous monitoring solutions to global rail networks.

Impacts & Benefits

1. Social Impact

Continuous real-time monitoring reduces derailment risks by catching defects early.
Improved passenger and staff safety, saving lives and avoiding major accidents.
Builds public confidence in the safety and reliability of Indian Railways.

2. Economic Impact

Low-cost prototype (~₹80k–₹100k) with high return on investment.
Predictive maintenance reduces inspection, repair, and failure-related costs.
Improved punctuality by minimizing unscheduled downtime and track closures.
Easy scalability across multiple coaches with minimal additional investment.

3. Environmental Impact

Energy-efficient IoT devices with battery and solar support help reduce carbon footprint.
Extended track life through early detection reduces frequent rail replacement and waste.
Lower fuel usage compared to heavy TRC operations for periodic inspections.

4. Technological Benefits

Advanced multi-sensor fusion (ultrasonic, accelerometer, vision, strain, temperature, GPS) for comprehensive diagnostics.
AI/ML-based analytics for defect classification, anomaly detection, and trend analysis.
Edge preprocessing for faster alerts with cloud backup for large-scale data storage.

5. Strategic & Operational Benefits

Indigenous solution that reduces dependence on imported TRCs and foreign spare parts.
Supports Atmanirbhar Bharat and Make in India initiatives.
Modular and upgradable design compatible with both broad and narrow gauges.
Any train can act as a monitoring train, enabling continuous track surveillance.
Seamless connectivity and user-friendly dashboards/mobile apps for railway staff.

Comparison: Existing Foreign ITMS vs Proposed Indigenous ITMS

Parameter	Existing Foreign ITMS (TRCs)	Proposed Indigenous ITMS (Retrofit Kits)
Deployment	Requires dedicated Track Recording Cars with limited availability.	Compact modular kit retrofitted under any train coach (broad and narrow gauge).
Cost	₹15–20 Cr per TRC with high operational costs.	~₹80k–₹100k per prototype, scalable and low operational cost.
Coverage	Limited runs; not all tracks are inspected frequently.	Continuous monitoring whenever the train runs on the track.
Technology Dependence	High dependence on foreign technology, imported spares, and support.	Fully indigenous system using locally available, efficient sensors and components.
Sensors	Focus on geometry, alignment, and ultrasonic with bulky hardware.	Multi-sensor fusion (Ultrasonic, Accelerometer, Camera, GPS, Strain, Temperature, etc.).
Communication	Offline data collection with manual upload and delayed analysis.	Real-time LoRa (rural) + Wi-Fi/4G (urban) connectivity and instant data transfer.
Data Analysis	Mostly offline processing; slower anomaly detection.	AI/ML-based cloud analytics with edge preprocessing for faster decisions.
Power Supply	High power requirement, dependent on train engine and heavy systems.	Energy-efficient design with battery backup and optional solar input.
Scalability	Hard to scale; limited number of TRCs per zone.	Highly scalable via multiple retrofit kits across trains and regions.
Maintenance	High maintenance due to complex, imported systems.	Easy maintenance with indigenous parts; local staff can be trained.
Environmental Impact	Higher fuel usage and periodic operation.	Efficient IoT devices extending track life and reducing replacements.
Social Impact	Slow response; derailments can occur between TRC runs.	Continuous monitoring with early warnings to save lives.
Strategic Value	Import-dependent with limited tech sovereignty.	Indigenous product enhancing strategic independence and innovation.

Skills & Technologies Highlighted

IoT & Embedded Systems Sensor Integration & Fusion Computer Vision (YOLO, CNNs, OpenCV) Edge AI on Jetson Nano Raspberry Pi / SBC Programming Python & ROS LoRa & 4G/Wi-Fi Communication Cloud Analytics & Dashboards Predictive Maintenance Railway Safety & Automation

Research & References

The solution is inspired and validated by prior research in track monitoring, structural health monitoring, and sensor-based diagnostics. Selected references:

Parkinson, H., & Iwnicki, S. D. (1999). An intelligent track monitoring system. Infrastructure Maintenance & Renewal.
Aw, E. S. (2004). Novel monitoring system to diagnose rail track foundation problems. Doctoral dissertation, MIT.
Wang, C. Y., Tsai, H. C., Chen, C. S., & Wang, H. L. (2011). Railway track performance monitoring and safety warning system. Journal of Performance of Constructed Facilities, 25(6), 577–586.
Redondi, A., Chirico, M., Borsani, L., Cesana, M., & Tagliasacchi, M. (2013). An integrated system based on wireless sensor networks for patient monitoring, localization and tracking. Ad Hoc Networks, 11(1), 39–53.
Imdad, F., Niaz, M. T., & Kim, H. S. (2015). Railway track structural health monitoring system. 2015 15th International Conference on Control, Automation and Systems (ICCAS), IEEE.
Amami, M. (2022). A Novel Design Concept of Cost-Effective Permanent Rail-Track Monitoring System. World Journal of Advanced Research and Reviews, 13(3), 451–473.