Smart Traffic Intersection Agent#
Note! This is a sample application intended for evaluation and development purposes only. For more information, refer to Intended Use
The Smart Traffic Intersection Agent demonstrates how Intel hardware performs in AI solutions for urban traffic management. It presents the use case of a comprehensive traffic analysis service that provides real-time intersection monitoring, analyzes directional traffic density, and provides traffic insights powered by a Vision Language Model (VLM).
The agent processes MQTT traffic data, manages camera images, and delivers intelligent traffic analysis through RESTful APIs. It supports sliding-window analysis, sustained traffic detection, and intelligent management of camera images to enhance traffic insights.
Here is its high-level architecture diagram, showcasing its core components and their interactions with external systems:
The Smart Traffic UI below shows how traffic and weather data is analyzed, and how summary and alerts are shown to the user:
Components#
The Smart Traffic Intersection stack includes the following containerized services:
MQTT Broker (Eclipse Mosquitto message broker) - Message broker for traffic data
DL Streamer Pipeline Server - Video analytics and AI inference
Scenescape Database - Configuration and metadata storage
Scenescape Web Server - Management interface
Scenescape Controller - Orchestration service
VLM OpenVINO Serving - Vision Language Model inference
Traffic Intelligence - Real-time traffic analysis with dual interface (API and UI)
Key Integration Points#
MQTT Communication: All services communicate via the shared MQTT broker
Docker Network: Services discover each other via Docker service names
Shared Secrets: TLS certificates and auth files mounted from
src/secrets/Persistent Storage: Traffic data stored in Docker volume
traffic-intelligence-dataHealth Monitoring: All services include health check endpoints