Ray Perception Sensor 3D Detectable Tags when self-playing

I'm doing a little experiment with multi-agent RL using self-play. I was wondering if my assumption about Ray Perception Sensor 3D Detectable Tags is correct. Basically both teams have the same goal, but both should distinguish their teammates and enemies through ray perception.

AFAIK detectable tags are one-hot encoded so if team1 agent looks like this:

And team2 looks like this:

to my best knowledge it should work just as I expect. Am i correct? Or maybe it's totally unnecesary and there is different solution for that case?

Thanks in advance. Have a nice day!


if you have the demos look at the soccer twos example