loader image
Top
ZIG PLATFORM · TECHNOLOGY

Inside ZIG AI.
How Verified Fare Evasion Detection Works.

ZIG AI connects what cameras detect with what the fare system actually recorded. Edge processing, AFC integration and journey intelligence work together so authorized teams receive a verified fare-compliance signal rather than a visual assumption.

ZIG_AI.PIPELINE
LIVE
EDGE AI
AFC
JOURNEY
ALERTS
LEARNING
boarding.detected cam_04 · 12:04:17
afc.transaction tap_ok - zone_B
journey.rule transfer_valid
signal.verified compliant
OVERVIEW

ZIG AI is part of the ZIG Platform. It extends ZIG validation and real-time intelligence to fare compliance by combining boarding detection, fare-transaction verification, configurable journey rules and operational reporting.

Architecture

Five Layers. One Verified Answer.

01

Edge AI fare verification

Processes time-sensitive validation close to the vehicle or station.

02

Computer vision + CCTV overlay

Detects boarding and provides authorized event context.

03

Open AFC integration

Confirms taps, payments and applicable fare records.

04

Journey intelligence

Applies transfers, fare caps, zones and concession rules.

05

Machine learning

Uses reviewed outcomes to improve deployment performance.

Edge
Back-office
Network
Detection Pipeline

From Boarding to Verified Alert

1
T+0MS

A boarding event is detected

2
T+120MS

Edge checks available fare transaction data

3
T+200MS

Journey rules reduce avoidable false positives

4
T+300MS

Verified anomaly is routed per agency policy

5
ASYNC

Reviewed outcomes support model improvement

Deployment

Edge-First by Design

Edge Processing (ZIG AI)

  • Boarding detection
  • Time-sensitive verification
  • Local alert generation
  • Continuity during network interruptions

Traditional Network Level

  • Journey rules
  • Fleet analytics
  • Synchronization
  • Model improvement
Method

Verified Detection vs. Visual Inference

Criteria Camera-only systems ZIG AI
Payment judgment Inferred from visible rider behavior Checked against AFC transaction records
Transfer rules Not evaluated Applied per agency configuration
Concessions & fare caps Often misread as evasion Recognized as valid fare conditions
False positives Higher, requiring manual review Reduced through verified data checks
Infrastructure Replaces or duplicates AFC Overlays existing AFC and validation
Ecosystem

What ZIG AI Connects To

CCTV & onboard cameras

Validators & fareboxes

Mobile ticketing & gates

AFC back-office

Operations dashboards

Enforcement devices

Learning Loop

Machine Learning That Improves With Deployment

01

Continuous improvement

Model updates informed by reviewed field outcomes.

02

Fleet-wide propagation

Validated updates distributed across the deployment.

03

Coverage analysis

Camera coverage and placement assessed over time.

04

Environment calibration

Tuned to vehicle layouts, lighting and boarding patterns.

Privacy

Non-Biometric by Design

  • No facial recognition or biometric matching
  • Behavior and fare status, not personal identity
  • Local processing for time-sensitive video analysis
  • Role-based access and audit controls
  • Agency-configured retention and image handling
  • No persistent cross-network movement tracking
PRIVACY.ARCHITECTURE

Vehicle / Station (Local)

Video
Detection
Verify

No biometric identifiers extracted or stored.

Controlled Event Output

Fare-compliance signal only • role-based access • audit logged

Deployment Model

Built to Overlay, Not Replace

Works with current AFC and validation infrastructure

Supports bus, BRT, rail, streetcar and open-platform environments

Scales from a defined pilot to broader rollout

Provides hotspot, trend and operational-impact analytics

DEPLOYMENT.SCALE
Pilot 1 route
Corridor 5 routes
Network full fleet
NEXT STEP

See the Architecture in Action.

Get a technical walkthrough of the detection pipeline, AFC integration, edge processing, journey rules and privacy controls with your agency infrastructure in mind.