What it contains
Positions, trajectories, lanes, movement zones, events and infrastructure state.
One system for sensing, understanding and operating traffic.
TeraFlow connects roadside sensors, 3D perception and tracking to show what is happening and how traffic is changing.
CAMERA + LIDAR
EDGE PROCESSING
3D TRACKING
See more than a camera can show.
Camera provides appearance. LiDAR provides geometry and depth.
TeraFlow combines camera and LiDAR into a traffic view that can be measured, replayed and reviewed.
TeraFlow ResearchPerception → Twin → Decision Support
REAL TIME
Continuous traffic-state awareness
SENSOR FUSION
Camera and LiDAR sensor fusion
3D TRACKING
Vehicles, pedestrians, cyclists and transit
EDGE AI
Local perception and event filtering
CONNECTED
Works with existing traffic systems
Understand movement without turning traffic monitoring into identity tracking.
TeraFlow can process perception at the edge, limit retention and provide operational views that focus on movement rather than identity.
- Edge processing
- Privacy zones · Configurable retention · 2D and 3D operational views
A live model of the road, not a static map.
TeraFlow continuously rebuilds road activity into a spatial traffic model.
Vehicles, motorcycles, pedestrians, lanes, movement and events stay connected across camera footage, 3D perception and historical playback.
Start with one intersection. Expand across the network.
- Multimodal detection and tracking for movement review
- Sensor layout defined per road environment
- Controller handoff scoped and tested separately
- Signal context
- Operational event review
- V2X handoff remains an explicit integration boundary
A traffic system should show what it sees, how conditions are changing and when a person needs to intervene.
TeraFlow engineering principle
TeraFlow Command
Monitor live traffic, inspect incidents and review site health across multiple locations.
Move between camera, 3D traffic context and recorded events without losing context.
- Live road and intersection views
- Camera + 3D traffic context
- Site and sensor health
- Incident review and playback
- Remote configuration
Bring us a road problem.
Traffic digital twin
A live model of road activity
The twin connects accepted observations into one spatial traffic state.
Accepted inputs
Low-confidence or rejected observations remain visible instead of being silently merged.
Continuous updates
The model reflects what selected sensors observe within documented coverage.
Zones and context
Operators configure the movements and events they need to monitor.
Playback
Review recorded scenes, trajectories and events with source and time retained.
Retention and access
Set data retention, processing location and authorized roles by project.
What it does not do
Signal control, V2X messages and physical actions require separate integrations.