Traf AI is a high-fidelity, deterministic traffic simulation platform designed for urban planning and corridor management. The system utilizes a stochastic flow engine and adaptive signaling intelligence to model complex intersection dynamics with a focus on throughput optimization and environmental impact analytics.
Built as a Digital Twin for urban infrastructure, Traf AI enables real-time scenario modeling, approach-based load balancing, and granular telemetry monitoring.
- Pulse Oscillation: Implements tidal inflow waves to simulate realistic peak-hour traffic patterns.
- Micro-Platooning: Generates vehicle clusters to mimic upstream signal phase transitions and platooning effects.
- Approach Commander: Cardinal-based control (North, South, East, West) for independent inflow rate and vehicle category calibration.
- Pressure-Weighted Scaling: Dynamically adjusts green-time (7s to 35s) based on real-time sensor data and lane occupancy.
- Safety-Red Clearance: Enforces a strict intersection clearing interval to prevent collision risk during high-speed transitions.
- Cyclic Phase integrity: Maintains a deterministic North-East-South-West rotation while allowing for proactive phase termination when queues are cleared.
- Telemetric Dashboard: Real-time tracking of throughput (Vehicles/Hour), Average Wait Time, and Level of Service (LOS).
- Sustainability Metrics: Dynamic calculation of total CO2 emissions and fuel efficiency based on idle-time vs. acceleration events.
- Driver Frustration Heatmap: Radial visualization of stagnant traffic for immediate bottleneck identification.
- Dynamic Diurnal Cycle: Seamless transitions between day and night lighting, affecting headlight visibility and ambient atmosphere.
- Weather Module: Support for Sunny, Rainy, and Stormy conditions with integrated physics adjustments for breaking distances and acceleration.
- Core: React, TypeScript, Vite
- Physics & Simulation: Custom Internal Engine (IDM-inspired logic)
- UI Architecture: Tailwind CSS, Shadcn/UI, Lucide Iconography
- State Management: Zustand
- Visualization: SVG-based Vector Engine, Recharts
- Node.js (v18+)
- Bun (Recommended) or NPM
-
Clone the repository:
git clone https://github.com/FTS18/AI-Traffic-Management-System.git
-
Install dependencies:
bun install
-
Start the development server:
bun dev
The system follows a reactive architecture where the simulation physics loop is decoupled from the UI rendering layer. The state is centrally managed in a performant Zustand store, allowing for flicker-free updates at 60 FPS across all telemetry components.
Developed as an institutional-grade simulation framework for modernizing traffic intelligence.