VAIC New Team Advice

I am a student at Colorado School of Mines, and I competed in VRC all four years of high school. I am looking to start a VAIC-U team at my school. Does anyone who participated in the pilot program have any advice?

Did you enjoy VAIC? What issues (besides COVID) did you face? What did you think of the provided software? Do you plan on competing again?

Thanks,
Nathan

3 Likes

I am no expert on vex ai, but you might want to read this thread.

@DRow maybe merge these?

1 Like

I have read through that one. The thread has some really good discussion about last season. I did not want to disrupt the thread with new team questions.

If DRow wants to merge, that is up to him.

1 Like

I’d just look at what the RECF puts out about next year before making your decision. The program will mature into something different as years go by. It is unclear exactly what vaic will look like next year. For the most part, it’s a good programming challenge. If all you want to do is program, I might suggest doing something else.

1 Like

Could you elaborate on this a bit? I am a MechE with decent experience in programming. That being said, I am primarily interested in the building challenge.

However, I am recruiting some of my CS friends. For a CS major at my school who only wants to work on machine learning/AI, would you recommend the competition?

It may be surprising but, as of last year vaic had no ai. If all you want to do is ai, I would recommend something else. With that, it is possible to deploy something like a neural network to detect objects with a vex robot. The caveat being you will have to start from scratch. This could require a team of dedicated developers. From a mechanical standpoint, vaic and vexU are identical.

1 Like

Due to lack of raw data, I assumed that it would be highly simplified and not Amazon-Robotics-caliber AI. Since the robots did not have AI, how would you describe the setup?

I’d say it’s more of a black box. They give you the stuff you need. Then you have to code how the robot interprets the data (x, y pos; heading; detected objects) from that black box.

1 Like

That makes sense. I expected that most teams would do something along those lines. Without a simulated field environment and loads of data, AI isn’t really possible. However, I hoped that incorporating machine learning/AI-like software would be more of a possibility. One of my friends has developed a variety of C++ neural nets that would be nice to play around with on a robot.

Do you plan on competing in vaic next year?

1 Like

This is one of my biggest issues.

For the most part, I like to write software more than build robots. Since vaic isn’t quite what I thought it was, I might try something else. Still, in following years it could turn into something much more.

1 Like