Group 5

Mike Estrada, Lu Yu, Ryan Liu, Jake Ramirez, Kevin Chen

Introduction

As countless companies attempt to have the first reliable autonomous car, our group attempted to find efficient ways to perform lower level autonomy tasks such as lane tracking and obstacle avoidance on a Traxxas RC car. The technology behind image processing is quickly getting better and easier so the demand for quick and novel computer vision algorithms is skyrocketing. We wanted to develop our own innovative vision systems that could run fast enough for a racing environment. This would be paired with a control barrier function (CBF), a special type of controller that ensures safe actions in safety critical situations and therefore prevents our car from potentially crashing into objects. We also wanted to implement a traction controller to maximize acceleration and grip in both straight line and high speed cornering situations.

Our goals ultimately were to:

1. Build an autonomous Traxxas RC car capable of navigating a track at high speed by using computer vision to detect lane markers and obstacles
2. Implement an MPC controller and control barrier function (CBF) to follow the track and avoid obstacles
3. Implement a traction control algorithm to maximize vehicle traction and acceleration