We started with an empty chassis to an autonomously moving robot that has the functionality of lane detection using OpenCV.

​

The original goal was to create a ChatGPT-controlled robot in a classroom setting, capable of tasks like navigating to a whiteboard and assisting students with problem-solving. While not all original goals were met, the team successfully enabled communication with the robot, generating paths and commands based on ChatGPT's responses to prompts.

​

Future goals include implementing fully autonomous navigation triggered by ChatGPT, improving path generation, and adding safety features like automatic stopping using LiDAR data. The project utilized OpenAI's ChatGPT API, DonkeyCar AI framework for path following, and ROS2 for command integration. Hardware components included a Jetson Nano, Oakd depth camera, ld06 LiDAR sensor, and point one Fusion Engine GPS.