Why our self driving racecar was DISQUALIFIED | MIT Driverless | Roborace Event 1.1 Rounds 3 & 4
This autonomous racecar is one of the biggest extracurricular projects I've ever worked on. So, when we were disqualified during our latest Roborace run on December 11th and 12th 2020, that didn't feel so great.
In the previous rounds, we fabricated "spline racer", an algorithm that near-optimally avoids obstacles and collects rewards. BUT we never got a chance to deploy it on the car... til now.
In round 3, there were connectivity issues with the car. We never got a chance to run our AI unfortunately. BUT things got a little spicy in round 4, when our obstacle avoidance/reward collection algorithm seemed to work super well, but one critical mistake left us out of the running for the top placements.
In this video, I'll talk about what happened, and what we're trying to do to fix this issue. This demonstrates how each team in the autonomous vehicle pipeline must work together to achieve the common goal. One part breaks, and something disappointing like this might happen... or worse.
Check out Roborace! / roborace
Feel free to leave any questions.
Please consider subscribing if you liked this video: https://www.youtube.com/c/ycubed?sub_...
Thanks for watching everyone!
~~~~~~~~~~~~~~~~~~~~~~~~
Follow me on Instagram: / kylieyying
Follow me on Twitter: / kylieyying
Check out my website: https://www.kylieying.com