- Artificial Intelligence
- Featured Case Studies
Transforming Toll Operations at Scale
Transforming Toll Operations at Scale
India's national highway toll network spans 1200+ plazas, generates over 20 million transactions daily, and runs on fragmented systems built by a dozen different vendors. Auriga IT consolidated that into a single real-time intelligence platform - and then extended it with a rules-based compensation management system that automates government scheme compliance for toll operators across the country.
India's national highways are managed by NHAI - the government body responsible for development, maintenance, and operations across hundreds of thousands of kilometers of road infrastructure. IHMCL, its subsidiary, specifically manages the FASTag-based national toll collection ecosystem. Together, they oversee 1200+ toll plazas that collectively process millions of transactions every day - making toll operations a mission-critical piece of national logistics and revenue governance.
Managing this at scale means coordinating equipment from multiple vendors, enforcing government mandates like 100% Electronic Toll Collection (ETC) lane uptime, and - most recently - automating compliance for passenger-facing government schemes that affect how toll operators are compensated.
A National Infrastructure Running on Disconnected Systems
Before Auriga IT's engagement, NHAI's operational picture was fragmented. Toll plazas ran systems deployed by different vendors, each with its own data model and reporting format. There was no single view of what was happening across the network - and when something went wrong, the first signal was usually a complaint, not an alert.
Toll Management Systems were deployed by multiple vendors across plazas with no central aggregation layer. Operational data lived in silos - there was no single dashboard showing network health, lane uptime, or revenue trends in real time.
Boom barrier failures, RFID scanner malfunctions, and ANPR camera outages often went undetected for hours. Issues were flagged by traffic complaints - not by the system. Each hour of undetected downtime meant revenue leakage and public grievances.
20 million daily transactions were generated across the network but never aggregated for analysis. Revenue trends, anomaly detection, and traffic forecasting were impossible without a centralized data layer that could read across all plazas simultaneously.
ANPR cameras and lane surveillance feeds generated terabytes of footage daily with no automated processing. Vehicle classification, lane misuse detection, and fraud identification required manual review - making consistent enforcement impossible at national scale.
Government mandates including 100% ETC lane uptime and scheme-based operator compensation required audit-ready compliance records. Without automated tracking, adherence depended on manual reporting from each plaza - slow, inconsistent, and difficult to verify.
When the Government introduced passenger-facing toll concession schemes, toll plaza operators needed a structured, formula-based compensation mechanism to replace the revenue they would otherwise have collected. No automated system existed to calculate or disburse these amounts.
The Toll Monitoring Control Center - TMCC
Auriga IT architected and delivered a modular, cloud-native platform called the Toll Monitoring Control Center (TMCC). The platform unifies IoT monitoring, data engineering, video analytics, and business intelligence into a single operational backbone for India's national highway network.
Industrial-grade IoT devices were deployed at each plaza to monitor boom barriers, RFID readers, CPUs, and ANPR cameras. Each device sends a continuous health ping to the cloud. A threshold-based alert system fires when any device exceeds 3 minutes of downtime - triggering escalation before a complaint reaches the helpdesk.
The result: maintenance teams are dispatched proactively. Equipment health is visible for every plaza, every lane, in real time - from a single dashboard.
A distributed data lake was designed to ingest live feeds from Toll Management Systems, NPCI (FASTag), traffic surveys, the Sukhad Yatra app, and mobile apps - simultaneously, at scale. Streaming ETL pipelines process over 20 million transactions per day with containerized microservices and autoscaling built in for peak-load tolerance.
The architecture is vendor-agnostic. Regardless of which TMS vendor deployed a given plaza, its data arrives in a unified format. One data lake. One schema. One truth across 1200+ locations.
ANPR camera and lane surveillance feeds were connected to GPU-enabled processing pipelines running deep learning models - YOLO and SSD architectures - for vehicle classification, counting, and anomaly detection. The pipeline handles over 10TB of video data daily in real time.
The system detects vehicle type mismatches, lane misuse, and suspicious behavior automatically - flagging incidents for review without human monitoring of footage. Classification results feed directly into the BI layer and compliance dashboards.
A custom BI layer was built with role-specific dashboards covering toll performance, equipment uptime, traffic volume, and revenue trends. Forecasting algorithms identify equipment failure risk, predict peak congestion windows, and flag revenue anomalies before they compound.
Every stakeholder level - from plaza operators to central ministry teams - has a dashboard built for their decision context. Data is no longer something you request from IT. It is something every decision-maker opens in the morning.
The entire platform powers a centralized Command and Control Center (CCC) that runs round the clock. Real-time dashboards for equipment health, lane uptime, and traffic movement are complemented by automated alert and escalation workflows tied to service SLAs - reducing both Mean Time to Detect (MTTD) and Mean Time to Resolve (MTTR) across all plazas.
Automating Operator Compensation Under Government Schemes
Toll plaza operators run on concession agreements. When a government scheme redirects revenue away from those operators, a compensation mechanism is required to settle the difference. The challenge was not the concept - the rules for compensation were defined by the Government. The challenge was turning those rules into a working, auditable, automated system that could run at scale across hundreds of plazas.
Auriga IT designed, developed, and deployed the compensation management system that does exactly that. The system reads scheme usage data from the TMCC data lake, applies eligibility rules, runs a formula-based calculation, and generates verified compensation outputs - with no manual reconciliation required.
How the System Works
Eligibility rules defined by the Government are encoded into the system. Each rule determines whether a given trip qualifies as a compensable event - based on vehicle class, pass type, and journey parameters.
The system tracks qualifying vehicle passages per plaza in real time. Pass holder journeys are distinguished from standard fare transactions automatically, with return journeys and repeat vehicle patterns accounted for in the logic.
Government rules specify trip categories that do not trigger compensation. These exclusions - including defined trip frequency limits - are enforced by the system. Operators cannot claim for trips that fall outside the scheme definition.
The formula accounts for the existing traffic baseline at each plaza. Compensation is calculated against the incremental impact of scheme usage - not against total traffic volume. This ensures operators are compensated for scheme-attributable revenue displacement only.
A structured formula computes the compensation amount per eligible trip, per plaza. The formula accounts for vehicle class, pass category, applicable fare, and qualifying journey count. Outputs are auditable and recomputable from raw event data.
The system generates compensation calculations on a defined cycle without manual input. Every output is traceable to the underlying trip data, eligibility decision, and formula applied - creating a full audit trail from scheme event to compensation figure.
Why This Engineering Problem Was Non-Trivial
The data inputs come from a live, high-volume stream - 20 million transactions per day across 1200+ plazas. Distinguishing scheme trips from regular traffic, applying eligibility rules correctly across vehicle classes, excluding non-qualifying journeys, and aggregating per-plaza compensation figures required both a solid data pipeline and precise rule encoding. Any error in the eligibility logic or formula produces incorrect operator payments at scale. The system had to be right, auditable, and defensible against review.
From Fragmentation to Full Visibility
| Area | Before TMCC | After Implementation |
|---|---|---|
| Operational Uptime | Reactive maintenance; frequent undetected lane outages | Threshold-based IoT alerts; proactive maintenance across all 1200+ plazas |
| Data Visibility | Transaction data locked in vendor silos; no cross-plaza view | Unified cloud data lake ingesting 20M+ transactions per day in real time |
| Decision Making | Manual Excel reports; days of lag before actionable insight | Live BI dashboards for instant decision support and predictive alerting |
| Video Intelligence | Raw footage stored; no automated classification or anomaly detection | AI pipeline classifies vehicles and flags lane misuse across 10TB+ daily video |
| Incident Response | Issues escalated after delay or via complaint | Auto-generated alerts reduce MTTD and MTTR across the network |
| Scheme Compliance | Manual tracking of operator compensation; no formula-based automation | Rules-based engine calculates and outputs operator compensation automatically, with full audit trail |
| Governance | Fragmented compliance reporting; difficult to audit across vendors | Real-time audit trails and automated compliance records across all plazas |
What This Project Required from Auriga IT
Apache Kafka and Spark Streaming pipelines handle 20M+ daily transactions with low latency. ETL processes cover deduplication, anomaly detection, and archival at national scale.
YOLO and SSD deep learning models run on GPU-enabled pipelines processing 10TB+ of daily video. Vehicle classification, counting, and anomaly detection run without human review of footage.
Containerized microservices on Kubernetes with autoscaling, load balancing, role-based access control, and 99.9% uptime SLA across cloud platforms (AWS, Azure, GCP).
Custom dashboards using Power BI, Tableau, and bespoke UI frameworks. Role-based visualizations, predictive models for equipment failure and congestion, and automated compliance reporting.
Designed and deployed a rules-based engine translating government scheme policy into automated computation - including eligibility logic, trip classification, exclusion rules, formula application, and audit-ready output generation.
Coordination with IHMCL, NHAI, and multiple empanelled TMS vendors across India. Solutions aligned with national mandates including 100% ETC lane uptime, FASTag adoption, and government scheme compliance requirements.
Core Technologies
Questions About This Engagement
What did Auriga IT build for NHAI?
How does the toll operator compensation system work?
What technology powers the NHAI government infrastructure data platform?
How many toll plazas does the platform cover across India?
Can Auriga IT build similar platforms for other government or infrastructure organizations?
What is Auriga IT's capability in government infrastructure data projects?
If your organization manages large volumes of IoT data, operational telemetry, or government scheme compliance - talk to Auriga IT about what the right data platform looks like for your context.
Talk to Auriga ITRelated content
Auriga: Leveling Up for Enterprise Growth!
Auriga’s journey began in 2010 crafting products for India’s [...]
