Linker Vision ← All products / Traffic

Traffic

Traffic Asset Digital Twin & Simulation

Using 3D Gaussian Splatting digital twins and NVIDIA Omniverse simulation, roads, bridges, tunnels and intersections are reconstructed in virtual space at centimetre-level precision and integrated with CCTV, VLM events, traffic and sensor data into a live state. In the twin, traffic simulation (SUMO), Omniverse and Cosmos rehearse crashes, congestion, flooding, roadworks and rare scenarios repeatedly so the AI learns to judge before it is used in the real world; the same twin serves CCTV placement optimisation, crash reconstruction, hazardous-intersection analysis, synthetic-data generation and road-safety education video. It is the training ground and decision base for transportation Physical AI.

Customer challenges

1

Roads, cameras and events have no shared 3D spatial reference

2

Signal, closure and detour decisions are executed without simulation

3

Congestion, evacuation and incident impact cannot be predicted before they happen

4

Crashes, flooding and rare hazards cannot be re-staged for AI training

5

Traditional 3D modelling is too slow and costly to keep a wide network current

Customer benefits

All transportation assets and events in one 3D space: from 'finding the feed' to 'seeing the state'

Decisions verified in the twin before execution; signal and guidance changes have an evidence base

Rare dangerous scenarios rehearsed safely thousands of times so the AI can judge before going live

CCTV additions and engineering investment land exactly on blind spots and risk points

Training-data generation efficiency rises sharply; adverse-weather recognition improves

One twin serves traffic control, maintenance, safety education and future autonomous testing; one investment, many uses

Key features

3DGS Digital Twin Reconstruction (Roads, Bridges, Tunnels, Intersections)

Reconstructs roads, bridges, tunnels, and intersections as 3DGS digital twins.

Traffic Simulation (Signal Timing, Lane Closure, Detour, Ramp Metering)

Simulates traffic under signal timing, lane closure, detour, and ramp metering changes.

Rare-Scenario Simulation (Rain, Flooding, Crash, Tanker Platoon)

Simulates rare scenarios such as rain, flooding, crashes, and tanker platoons.

Synthetic Data Generation (Adverse Weather, Night, Tunnel)

Generates synthetic data for adverse weather, night, and tunnel conditions.

CCTV Placement Optimization & Blind-Spot Analysis

Optimizes CCTV placement and analyzes blind spots.

Crash Reconstruction & Hazardous Intersection Analysis

Reconstructs crashes and analyzes hazardous intersections.

Crowd & Evacuation Simulation

Simulates crowd movement and evacuation.

Before & after

Before Wide-area 3D modelling takes months After — with Linker AI Nexa About 4 weeks from scan to live, modelling lead time -75%
Before Model accuracy unverified After — with Linker AI Nexa Measurement error ≤ 2%, ≥ 90% of survey points pass
Before Rendering too slow for live operational use After — with Linker AI Nexa Real-time above 60 FPS, 10–100x faster than NeRF
Before Control changes tested only in the real world After — with Linker AI Nexa Verified in simulation first, -15 minutes congestion around an industrial park
Before Evacuation plans untested After — with Linker AI Nexa Peak-platform evacuation time -8%, evacuation capacity +20%
Before Rare hazards observed a few times a year After — with Linker AI Nexa Thousands of synthetic rehearsals per scenario