System: OPERATIONAL
Active Site: Baspani Yard, SE Railway
Accuracy: 98.7%
Scan Cycle: <3 sec / wagon
LiDAR Precision: ±2mm
Uptime: 24 / 7
Live Deployed · Indian Railways

IWLARS
PLATFORM

Intelligent Wagon Load Analysis & Reporting System

AI-powered LiDAR-based 3D load distribution analysis and automated visual inspection for freight wagons — preventing derailments, ensuring compliance, and eliminating manual yard bottlenecks at Indian Railways loading yards.

98.7%
AI Accuracy
<3s
Scan Cycle
±2mm
LiDAR Precision
24/7
Operation
View Live Deployment → Technical Overview
SYSTEM OVERVIEW — IWLARS · REV 2.4 LIVE OVERHEAD GUIDE RAIL / CABLE TROLLEY UNIT LiDAR 3D LiDAR SENSOR ±2mm · 1M+ pts/scan WGN-4471 · COAL · SE RAILWAY 9.0m WAGON LENGTH SCAN ZONE EDGE AI SERVER CNN · <3s inference → DASHBOARD REPORT
Problem Statement

WHY MANUAL WAGON
INSPECTION FAILS

Traditional visual inspection of loaded freight wagons is error-prone, slow, and incapable of detecting load imbalances that cause derailments, track damage, and billion-rupee losses across India's freight network.

🚨
Derailment Risk from Load Imbalance
₹200Cr+
Asymmetric coal/ore loads shift the wagon's centre of gravity, dramatically increasing derailment probability at curves, gradients, and high-speed sections. A 15%+ lateral imbalance is considered critical.
Safety Critical
⚙️
Accelerated Axle & Track Wear
40%
Uneven loading accelerates axle fatigue, bogie stress, and ballast degradation. Unbalanced loads cause up to 40% faster infrastructure wear versus properly loaded wagons on the same route.
Infrastructure Damage
🕐
Manual Inspection Bottlenecks
~8min
A single manual inspector takes 7–10 minutes per wagon — creating serious throughput bottlenecks at high-volume loading yards. A 58-wagon rake requires over 8 hours for full manual inspection coverage.
Operational Inefficiency
📋
Zero Digital Audit Trail
0%
Paper-based inspection records offer no tamper-proof audit trail, no analytics capability, and no incident correlation. Compliance reporting requires manual data compilation — prone to errors and delay.
Compliance Risk
Technology Pipeline

HOW IWLARS WORKS

A 4-stage integrated pipeline from physical LiDAR acquisition to digital intelligence — completed in under 3 seconds per wagon.

SYSTEM PIPELINE — END TO END · IWLARS v2.4
STAGE 01 🚂 Wagon Entry Trigger sensors detect wagon entry into scan corridor zone Proximity · RFID Tag STAGE 02 📡 LiDAR Point Cloud Trolley traverses wagon. 1M+ 3D data points captured at ±2mm res. 3D SLAM · Multi-return STAGE 03 🧠 CNN AI Analysis Load volume, balance ratio & anomaly scored on edge GPU hardware ResNet · CUDA · Edge STAGE 04 📊 Dashboard Report Live metrics, 3D profile, pass/fail verdict & automated PDF report Web Dashboard · API · Alerts TOTAL CYCLE TIME: < 3 SECONDS PER WAGON · 24/7 AUTONOMOUS OPERATION · ZERO CLOUD DEPENDENCY
Load Distribution Intelligence

3D LOAD PROFILE &
WEIGHT DISTRIBUTION

TOP-DOWN LOAD HEATMAP — WGN-4471 · COAL BOGIE A BOGIE B ← WAGON LENGTH AXIS → HOT ZONE 68% MID ZONE 24% LEFT SIDE: 52% RIGHT SIDE: 48% LOW HIGH LOAD DENSITY SCALE
AI-Computed Weight Distribution
The 3D LiDAR point cloud is processed by Aerovania's CNN model to generate a real-time top-down heatmap of cargo density across the wagon floor. This identifies imbalance zones, hotspots, and unsafe distribution patterns before dispatch.
96.2%
Load Balance Score
52/48
L/R Weight Ratio
82.4%
Fill Level
PASS
Dispatch Status
Dispatch Thresholds
Balance ≥ 90% — CLEARED FOR DISPATCH
Balance 75–89% — REBALANCING RECOMMENDED
Balance < 75% — HOLD — DO NOT DISPATCH
System Architecture

FULL TECHNOLOGY STACK

Every layer of IWLARS is engineered for industrial-grade reliability — from the physical sensing hardware to the edge AI inference engine and operator dashboard.

SENSING
Pulley-Guided LiDAR Trolley System
Motorised overhead cable-guided trolley with 3D LiDAR sensor payload. Traverses full wagon length automatically. Multi-return LiDAR generating 1M+ points per scan at ±2mm spatial accuracy.
±2mm resolution1M+ pts/scan360° coverage
HARDWARE
VISION
Computer Vision Camera Array
High-resolution industrial cameras mounted alongside LiDAR for RGB visual inspection — detecting wagon body damage, cracks, open hatches, spillage, and coupling anomalies.
4K industrial camsStructured lightIR night mode
CV
DATA
3D Point Cloud Processing Engine
Real-time SLAM-based georeferenced point cloud reconstruction. Sensor fusion pipeline integrates LiDAR + camera data. Outputs volumetric wagon interior model for downstream AI analysis.
SLAM positioningSensor fusionPCL library
PIPELINE
AI ENGINE
CNN Load Analysis & Inspection Model
Deep convolutional neural network trained on annotated coal/iron ore load datasets. Computes 3D load heatmap, lateral balance ratio, fill percentage, and anomaly confidence scores with 98.7% accuracy.
ResNet backboneCUDA accelerated98.7% accuracy
AI / ML
COMPUTE
On-Premise Edge AI Server
All inference executes on site-deployed edge hardware — no cloud latency, no connectivity dependency. NVIDIA GPU-accelerated server in ruggedised industrial enclosure. Results delivered in <3 seconds.
NVIDIA GPU<3s end-to-endIP55 enclosure
EDGE
OUTPUT
Web Dashboard + Reporting + API
Real-time operator dashboard with 3D load visualisation, historical analytics, automated PDF report generation per wagon, SMS/email alerts for failures, and REST API for SCADA/ERP/TMS integration.
REST APISCADA readyAuto PDF reports
SOFTWARE
📡 LiDAR Hardware Specifications
Sensor Type: 3D Multi-return Time-of-Flight LiDAR
Spatial Resolution: ±2mm
Point Density: 1,000,000+ points per full scan
Scan Range: 0.5m – 30m operational
Coverage Pattern: 360° horizontal, 120° vertical
Mounting: Motorised pulley-guided overhead trolley
IP Rating: IP65 dust/moisture proof for yard environment
VLP-32C compatible100Hz refreshIP65 rated
🧠 AI Model Parameters
Architecture: Custom ResNet-based 3D CNN
Training Data: 10,000+ annotated wagon scans
Commodities: Coal, iron ore, aggregate, ballast
Outputs: Load balance %, heatmap, fill level, anomaly flag
Inference Time: ~800ms on edge GPU
Accuracy: 98.7% on held-out test set
Model Updates: Remote OTA with yard-specific retraining support
PyTorch frameworkTensorRT optimisedOTA updates
🔗 Integration & Compliance
API Protocol: REST / JSON + Webhook callbacks
SCADA Integration: Modbus TCP, OPC-UA support
Railways TMS: Compatible with Indian Railways FOIS
Data Storage: On-premise SQL + file archive
Report Format: PDF/A, CSV, JSON
Security: Role-based access, audit log, TLS encryption
Compliance: Indian Railways load inspection standards
IR FOIS readyOPC-UATLS 1.3
Platform Capabilities

SIX CORE CAPABILITIES

IWLARS delivers a complete inspection and analytics suite — from raw sensor acquisition to automated compliance documentation.

📡
3D Load Profiling
Full volumetric 3D model of cargo inside the wagon — showing surface topology, fill depth map, and mass concentration zones across the entire wagon interior.
  • ±2mm spatial resolution
  • Full interior coverage map
  • Surface topology extraction
  • Volumetric fill calculation
⚖️
Weight Distribution Analysis
AI-computed lateral and longitudinal weight distribution — quantifying imbalance percentages across bogie zones and flagging wagons exceeding safe dispatch thresholds.
  • L/R balance ratio computation
  • Front/rear distribution check
  • Dynamic safety thresholds
  • Derailment risk scoring
👁️
Automated Visual Inspection
Computer vision scan of wagon body detecting structural damage, cracks, open hatches, coupling anomalies, and cargo spillage — all classified by AI without human involvement.
  • Crack & damage detection
  • Hatch / door status
  • Spillage & overfill flag
  • Coupling condition check
⚡
Edge AI Processing
Entire inference pipeline runs on-premise edge hardware — zero cloud dependency, sub-3-second results, and full operational continuity in offline or low-connectivity yard environments.
  • <3 second total cycle
  • No internet required
  • NVIDIA GPU accelerated
  • Fault-tolerant architecture
📊
Automated Reporting
Every scan auto-generates a structured digital record with load metrics, 3D heatmap, pass/fail verdict, timestamp, and wagon ID — archived and accessible for compliance audits.
  • Per-wagon PDF reports
  • Shift & rake summaries
  • SMS / email alerts
  • Historical trend analytics
🔗
SCADA / ERP Integration
REST API and webhook interfaces for integration with existing Railway SCADA, FOIS, ERP, and freight management platforms — zero disruption to existing yard operations during deployment.
  • REST API + Webhooks
  • Modbus TCP / OPC-UA
  • IR FOIS compatible
  • Custom integration support
Live Reference Project

DEPLOYED & OPERATIONAL
IN THE FIELD

IWLARS is not a concept — it is a live, commissioned system running at South Eastern Railway's Baspani Yard, CKP Division.

South Eastern Railway
CKP Division
AI-Based Automated Visual Wagon Inspection & Load Analysis System — Baspani Yard, Jharkhand
Deployment Type
Turnkey Installation
Commissioned
2025
Commodity
Coal / Iron Ore
Client Type
Indian Railways
Project Scope
▸Installation of pulley-guided LiDAR trolley infrastructure at Baspani Yard loading corridor
▸Custom CNN model training on coal/iron ore wagon load patterns and visual anomaly datasets
▸On-premise edge AI server deployment with full redundancy and IP55 enclosure
▸Real-time operator dashboard with automated PDF reporting and alert system
▸Yard management workflow integration and operator training programme
98.7%
AI Detection Accuracy
<3s
Scan Cycle Time
±2mm
LiDAR Precision
24/7
Operational Uptime
CROSS-SECTION VIEW — BASPANI YARD INSTALLATION TROLLEY LiDAR
Target Applications

DEPLOYABLE ACROSS
INDIA'S FREIGHT NETWORK

🏭
Coal Loading Yards
Thermal power plant sidings, Coal India (WCL, CCL, SECL) loading terminals with high daily wagon throughput.
Coal India · Railways
⛏️
Iron Ore & Mining Sidings
NMDC, Tata Steel, and JSW steel plant rail sidings handling iron ore and bulk mineral freight dispatches.
Mining · Steel
🚉
Marshalling Yards
Indian Railways formation yards requiring wagon load verification before rake assembly and train dispatch clearance.
Indian Railways
🏗️
Infrastructure Projects
NHAI and large infrastructure corridors using material transport wagons for ballast, aggregate, and construction supply.
NHAI · NHPC
Aerovania
Contact Us