Over the summer I participated in the University of Arizona’s CAT Vehicle REU funded by the National Science Foundation. During this 10 week experience, I and two other students worked with our graduate mentor to develop a lane detection and following system for the University’s Cognitive Autonomous Testing (CAT) Vehicle.

The CAT Vehicle is a hybrid Ford Escape configured for drive-by-wire and containing the neccessary computer systems to control the car using the Robotics Operating System.

My team and I developed a controller for the car to follow lanes using input from a front facing camera combined with GPS odometry. Along the way, we created a new way to model the path of the lane by improving upon an existing sliding window algorithm. We submitted a short paper detailing this work to the IEEE IRC 2019 conference (Github repo with paper and code).

Below is a short video to describe the entirety of my team’s work.