Video restuls seems to show that it's using baseline controller

[Na-Teng Hung]

Hello, I checked both of our teams videos, and they all look like it's running the baseline controller...despite that we submitted ours with our custom one...

Could you verify if we made a mistake somewhere?

[Na-Teng Hung]

And interestingly enough, while I was comparing side by side the baseline controller...

It seems that there are some extra random penalties despite hitting the pedestrian at the exact same time, and exact angle?

(left score 151, right score 140). Since we got 10 points extra, I believe it's part of the deviation penalty category, which doesn't make sense because the car is still within the track limit.

[Generalized RAcing Intelligence Competition]

Hi Competitors,

We did use your submitted controllers. You submitted a single python file that use (scipy + matplotlib), right? Although I didn't look into details, it seems like both your controllers are building on the baseline controller.

Regarding the lane invasion penalty, I've checked the log for both teams, it's actually not the exact same time and exact angle, although that seems like it. The penalty is assigned because, when you collide, we will forcefully reset your position, and during the reset process, your car is temporarily out of bounds. We will decide how to deal with this situation today. Thanks for bring it up.

Bests,

Yan

[Na-Teng Hung]

Thank you. We'll check our code again. Appreciate it.

[Na-Teng Hung]

Just curious, what happens if your end detects that our controller has a code error? Does it resort to using the baseline controller?

[Generalized RAcing Intelligence Competition]

Currently we do not plan to reveal the detailed architecture of the testing pipeline for fairness reason (at least not releasing before the competition deadline)

But you raised an interesting point. Let me change some configurations and run your code again in the next couple hours.