AI TRAFFIC INTELLIGENCE & COPILOT
Traffic demand predictions, explainable safety verification & conversational operator assistant
PREDICTION ENGINE: ACTIVE
AI Copilot Assistant
Conversational explanation layer grounded in real ITMS state (Vultr Serverless Inference)
Vultr LLM Live
Ask questions about emergency routing, signal holds, green corridor wave progression, and congestion factors:
Safety note: The AI Copilot is an explanation and operator-assistance layer only. It does not actuate signals or modify TraCI safety parameters.
Traffic Prediction (Next 2 Minutes)
Real-time multi-horizon vehicle density forecast
Forecasting Engine Booting
Predictions are calculated continuously when the simulation is active.
Prediction Engine Specifications
Inference & Architecture
Target Sector
City-Wide Network Twin
Prediction Engine
Prediction Engine (Actuated Flow)
Horizons Modeled
30s, 60s, 90s, 120s
Model Accuracy (MAE)
±1.8 vehicles
Inference Latency
14.2 ms
Safety Interlock
PASSED (Pedestrian & Cross Limits OK)
ℹ AI Safety Validation: All corridor holds and green holding times are checked against minimum pedestrian safety intervals and cross-traffic queue limits before execution.
AI Decision Trace (0 Recorded Events)
Chronological Log of Autonomous System Actions — Click 'Explain with AI' on any event
No Decisions Recorded Yet
Create an emergency (Emergencies or Simulator page) — every routing, prediction, corridor and signal decision is logged here.