The Powder Toy

TPT evolution test2 - Schmolendevice by Schmolendevice

Schmolendevice
Schmolendevice
27 / 2
25th Aug 2015
14th Jul 2016
Schmolendevice's version with an 8 neuron 'signal filter' ensuring that at any time, only the 'most often queried' direction of movement is taken. This version will incorporate simple plnt seeking and obstacle avoiding neuronal functions.

Comments

  • Zamal
    Zamal     13th Jul 2023
    yo where did u learn that stuff???
  • QuanTech
    QuanTech     4th Jul 2016
    wwwwwwwwwow!
  • yashar1995
    yashar1995     22nd Dec 2015
    neural networks!!
  • Karakanlud
    Karakanlud     21st Sep 2015
    I'm working on adding action potential function to my neurons, take a look! id:1857665
  • Karakanlud
    Karakanlud     21st Sep 2015
    very nice, +1
  • perogiepro
    perogiepro     26th Aug 2015
    oh wow i see now how it really is thinking: "wait, I went up and found plnt before" so if I when down I should go up to find plnt. Its really calculationg. each direction memorized adds together ass a mini memeory thing. we really need to implement this thing in. Plus it doesnt overload with too much plnt (for when it can replicate) . Im going to add it.
  • Schmolendevice
    Schmolendevice     26th Aug 2015
    You have to place brick within 9 or 10 pixels of it and it will try to avoid.
  • Schmolendevice
    Schmolendevice     26th Aug 2015
    I'm getting there. Right now the only issue is that the motor outputs are going to the wrong 'limbs' plus I need to start it off with an initial direction of movement.
  • perogiepro
    perogiepro     26th Aug 2015
    i see where your going but it falls apart easily and isnt random (TO START) so it rams into the water each time and eats itself with cray. but you made this so you know how it works so try to fix this for the test. you so far have put the most work into this project so amzing job.