scroll to explore

Traffic Optimization Simulator

A 4-way intersection traffic simulation comparing fixed-time signal cycles against adaptive, intelligent control. Built with a FastAPI backend and live dashboard, the system integrates YOLOv8 computer vision for real-time directional vehicle detection with a hybrid Deep Q-Network (DQN) and Genetic Algorithm (GA) engine to dynamically calculate optimal green-light durations based on traffic queue lengths.

The hybrid control algorithm prioritizes high-congestion approaches while enforcing maximum wait bounds to prevent starvation across all lanes. The project was officially awarded a Government of India Copyright Registration (Cert. No. SW-2026023150).

View on GitHub

Services

Deep Q-Network (DQN) & Genetic Algorithm
Computer Vision (YOLOv8)
Adaptive Traffic Control Systems
FastAPI Backend & Simulation Design

View on GitHub

InvisiMail

Have a project or engineering role in mind?
CONTINUE TO SCROLL