VEX AI - Can the Nvidia Jetson Xavier NX be used instead of the Nvidia Jetson Nano for Vex AI?

I clarify that not to participate in the competitions, but to learn the programming code.jetson-xavier-nx-dev-kit-2c50-L

Which programming code ? VEX provided AI software needs a hardware dongle to run.


It is not yet clear how they are programmed. The Jetson Xavier NX development board is more powerful than the Jetson Nano, I suppose it can work on either one, but both are designed to install Linux Ubuntu on them. So, the software that Vex will provide and that is not yet known, will run on Ubuntu OS?

If you have registered for the AI competition a Jetson Nano image is provided, it has the VEX software pre installed and requires a hardware dongle to operate. I’m still not sure what you are asking, if you are registered you could try and run the image on a Xavier, but I have no idea if it would work. If you are not registered for the AI competition then you have no way to run the VEX software.


I understand, what happens is that I bought a Jetson Xavier NX. My question is if I can use the software without being registered, since I have the hardware and the Vex V5 system. Thanks

No you can not. As I already said.
VEX provided AI software needs a hardware dongle to run


Look for freely available and open-source alternatives instead (OpenCV, etc.).

The VEX AI software, as jpearman has stated, is not available to anyone that has not registered for VAIC.


:sob: :sob: :sob: :sob:


I’d suggest returning the Xavier, since, iirc, that is far more expensive than the Nano, anyway.

No, I didn’t buy it to use with Vex. In fact, I bought it because I wanted to venture into AI for other types of projects, it also offers me other types of services.

1 Like

Also its wayyy more powerful and power efficient than a Jetson Nano. I think its a totally reasonable platform to do AI on, given you are willing to all of vex AI yourself without any help from VEX.


This is a pretty good write up about the differences between the two for people that want to roll their own AI system.

I don’t have a dog in this fight, and I’m not doing VEX AI. I’m seeing that VEX is providing a consumer grade AI appliance, vs roll your own. If you are going to do VEX size robots with one / two cameras then Nano will fly. Power is the big factor with the NX eating watts like I eat M&Ms. But if you have bigger goals and the extra $200 doesn’t make you twitch, then NX is you.

This is a demo year, no idea what 2021-22 will look like. Your cool code from the NX should back port, but based on “needs a dongle” comment it’s not clear you can do whatever software you want.

I’d suggest visiting Chief Delphi and search for Jetson and vision threads, that may help your decision and coding process better.

Good luck


it’s right, the characteristics are much superior and it allows to install an M.2 SSD, while the Nano does not.

1 Like

NVIDIA has provided a large amount of tools to assist in developing robotic systems. My personal recommendation is the Isaac SDK, I’ve worked with it for about 6 months now and it is quite nice and comprehensive(also free I believe, but don’t fully quote me on that) . However I wouldn’t jump into the Isaac SDK without some experience as it gets quite overwhelming at times. I am a personal fan of the Xavier models, but they are definitely overkill for a controlled environment like the VAIC game. I wouldn’t rush into buying one if you have no experience in any ML or CV, or if you don’t need that much computing power. Also can confirm Foster’s point about the higher-end NVIDIA boards being incredibly power hungry, they are deceptive.

Edit for clarity: I do not work, and have never worked with the NX board. I will be experimenting with the AGX model from here on out, and have worked extensively with the TX2.