Pygame / Driverless Car AI
Driverless Car Simulator
Drivlerss Car simulation made using pygame. This project incorporates a use of A* and D*Lite pathfinding algorithms as well as neural networks trained using reinforcement learning.
Overview
This project was developed as part of a larger AI-driven traffic simulation system. It was created to explore and better understand pathfinding algorithms and reinforcement learning techniques in autonomous navigation. The simulation includes systems that allow vehicles to efficiently navigate environments, dynamically recalculate routes around unforeseen obstacles, and perform collision-avoidance manoeuvres to safely dodge other cars in real time.
Features
- • A* path finding
- • D*Lite Obstacle avoidance
- • real time responses to traffic lights
- • CNN neural networks used to avoid on-coming traffic
Challenges
One of the main challenges was implementing the D*Lite algorithm for dynamic pathfinding. This required careful handling of edge cases and ensuring that the algorithm could efficiently update paths in response to changing obstacles. Additionally, training the reinforcement learning model to effectively navigate traffic scenarios was a complex task that involved tuning hyperparameters and designing a suitable reward structure.
Tech Stack
Links
Coming soon!
Gallery
What's New (17-5-2026)
Latest Update
Date: 17-5-2026
Added to Site
This project made before the website was built, so there are no updates to show. However, I am planning to add a video walkthrough of the project in the near future, so stay tuned!
Development Journey
Project Timeline
17 May 2026
This project was added to the website
This project is still in development, but I wanted to add it to the website early to show the progress. Stay tuned for more updates and a full project walkthrough video in the near future!