And maybe replacing the chaos cores with action potential neurons, so it would not give random signals.
The next three planned updates are a deciphering mech. for the memory unit, so the creature can have different profiles, like, it can reprogram itself to actually stick to the walls instead of avoiding them, a mapping mechanism, which would add a sequence of actions executed after taking some steps, and a two level decision system, which would make the brain able to control lesser functions (e.g. memory/relaxed signal decoder reprogramming) by itself.
NorthMustang: Another problem is how the distance information could be stored.I think the first method could be a sequence of actions, when the creature stops and gathers information about the environment, instead of doing it continuosly.
NorthMustang: I haven't found one yet. I've thought about using something like heat, but that would be very unreliable. I've thought about PHOT scattering, that might provide exact distanced.
You could've had long range detectors that told the bot where the walls were... Of course it would be too tedious to make; Unless of course you had a form of long-range, high-speed, no-damage detectors that opened up the possibility to the bot that it could go that direction... And then it could detect what element it's near, and further base its decisions off of that.
i have no idea what any of this does... +1
+1 you have successfully created a roomba 980 robot vaccum cleaner.