2D Autonomous Driving Project

the 2D Autonomous driving project was completed in CS271, Intro to Artifical Intelligence using Anaconda, gymnasium and pygame

● Trained an agent using a PPO and DQN model that is able to navigate a car from a starting point to a goal destination in an efficient manner, without veering off the road.

● Compared the different pros and cons between both models with similar amounts of training to find that PPO was 14% more successful

● Iterated on top of an existing Gymnasium environment meant to train race cars. We took the existing continuous and discrete input systems, completely revamped the environment and trained on that.