The International Autonomous Robot Racing Competition (IARRC) involves racing a small autonomous robot on a track in two different scenarios: Drag race and Circuit race. Both races take place on a pavement surface, and the vehicle must remain within boundaries designated by cones and lines. The drag race occurs on a straight track, where the objective is to drive to the end of the track as fast as possible while also stopping within some distance after the finish line. The circuit race requires the robot to make turns on a significantly longer track.
During Summer 2016, I traveled to the University of Waterloo with several RoboJackets teammates to participate in IARRC. I was only brought onto the IARRC team just a few weeks before competition. In that short time, I wrote the perception and planning software that would eventually help the team win 1st place overall for the competition.
Our robot has an Intel Nuc, and a Logitech webcam as its only sensor. The camera identifies the lane boundaries and cones using simple color thresholding, and steers toward the direction with the lowest cost (least obstacles).
Our vehicle completing one lap of the circuit course.