Developed an intelligent traffic management system with real-time vehicle detection and dynamic light control using AWS and simulation tools. Key components include:
- Simulated traffic with SUMO on AWS EC2 and streamed screenshots every 5 simulation steps via Flask and TraCI.
- Used AWS Kinesis for real-time image streaming; images processed with OpenCV in AWS Lambda to estimate vehicle counts.
- Triggered Lambda functions to update traffic lights based on analysis, applying decisions live in the simulation via Flask.
- Python
- Flask
- SUMO
- TraCI
- OpenCV
- AWS EC2
- AWS Lambda
- AWS Kinesis
- DynamoDB
- Terraform
The infrastructure is provisioned using Terraform, which automates the following resources:
- Creation of a VPC to isolate the environment.
- Provisioning of an EC2 instance to host the SUMO simulation and Flask application.
- Configuration of AWS Lambda functions for image processing and traffic light control.
- Use of AWS Kinesis for real-time image streaming.
- Storage of Terraform backend state securely in an S3 bucket with state locking enabled.
- Setup of DynamoDB tables for storing vehicle count and traffic signal data.
- Configuration of IAM roles and policies to securely grant necessary permissions to Lambda functions, EC2 instances, and Kinesis streams.
This setup ensures a scalable, secure, and reproducible deployment environment.
Below is the architecture diagram illustrating the components and data flow of the intelligent traffic management platform:
