This gym environment models an autonomous vehicle driving scenario.

The goal is to reach the green rectangle. Obstacles are shown by red rectangles. The autonomous vehicle is shown by the blue rectangle. The arrows in the upper right corner show the front a rear wheel steering angles. Longitudinal velocity is fixed.

The environment was evaluated with a modified version of TD3 (twin delayed deep deterministic policy gradients). Instructions for training on the environment are provide in the Readme: TD3 Repository

A presentation of the project's goals and lessons learned

Slides

Final Report