Indie game storeFree gamesFun gamesHorror games
Game developmentAssetsComics
SalesBundles
Jobs
Tags
(2 edits) (+2)

SUGGESTION TO IMPROVE THE A.I. to be a better runner

Like in real life, babies don't start trying to run, they

try first be stand up.

The program should divide the learn moves into 3 steps to lear how to run:

1- STAND UP : this is important to software knowing what is the correct creature  position, to kwnon what is theyer foot, legs, head etc. 
in this process at some moment we should to able  to use a drag force using the mouse cursor, to try desestabilish the creature, or thrown heavy objects in it direction to test and see how creature try stand up with "stranges" forces on it body.

2- WALK: After learned how stay on foot, should now learn how to walk, becouse have a huge diference between walk and run, some animals (like a cat) when they are running, some times (mileseconds) they dont touch the ground and walking have slower moves and software should try to find the best way to consuming less energy in muscles (becouse walking is the most eficient way to move, differently of running, becouse is more explosive moves and higt energy musscle consuption..

3- finally using all the informations from step 1 and 2 to  learn how to run, tryng to be the faster as can,


thats it THANKS.


(+1)

But how would the game know what was your legs, and what was your body?

I usually make hyper-symmetrical creatures actively intended to move by rolling; having the simulation make them 'stand' first would mean the simulator would never find the intended movement pattern for my 4-fold mirror rotationally symmetric creations meant to roll. Granted, I realize I am likely an outlier among the playerbase; for non-rolling creations, your suggestion does make sense. If it were something that could be toggled, I'd be fine with it; just as long as it isn't forced on me when trying to make and evolve a rolling creation.

(+1)

you aren't an outlier, i'm currently doing this.