I had a great time too. I didn't make it around the course in the competition, as I was having calibration issues with my IMU. However, I did make it to the hay bales during practice. I have a feeling the rocky ground messed with my IMU, but further testing is needed.
Interestingly enough, I had more fun on Friday than I did on Saturday. On Friday, I'd perform a run, identify a problem, fix it, and run it again. As time passed, my car got better and better. On Saturday, my first heat was a disaster. I knew how to fix it, but it required remapping the course, which I couldn't do, so it was a bit frustrating.
As this was my first year, it was interesting to see everyone's robots, their design ideas, and to learn that no one had 100% confidence that they'd complete the course (with the exception of one lady, and their robot didn't score any points).
I didn't get to look at the winner's robot, but I'm surprised that it used image processing. From the little I've used with openCV, I would think that it would be tough to distinguish between the brown, textured hay and the brown, textured ground. That Intel camera might have some really good processing on it though.
I am Groot was using a Neo-8N GPS, and claimed he could get around .5 meter accuracy with it. Considering he made it through the bales with that (no other obstacle avoidance sensors), that's damn accurate.
Another guy that completed was using some Traxxas chassis (I don't know much about RC cars) and the stock motor with the XL5 esc, which is exactly what I'm using. I was surprised how slow he got the car to move by the barrels without stalling the motor, as I couldn't get my car to go that slow without stalling. He did it by varying the PWM cycle above and below the stall speed, something I hadn't thought about. He was also using dead-reckoning, using the standard RPM sensor on the gearbox and the gyroscope, which he claimed was way more accurate than trying to use the mag sensor. Last of all, they had a pretty nice lidar mounted near the rear of the vehicle for obstacle avoidance. I don't remember the model of the lidar, but they said it was their "low-end" model that cost about $1000.
Props to Ted's bumper design and using flexible filament. I hadn't thought of that, and may try it next year. I had 2 3D printed bumpers...both shattered when I hit an obstacle, and I spent much of my time repairing them. Even after repair, one of them was severely warped, probably because I left it in the sun.
I'd love to go again next year.