[ODE] Community page

[Mattias Fagerlund]

> Pretty interesting stuff.
> I'm doing similar things with a two legged runner and using Mathew
> Wall's GALIB.

My current implementation evolve NNs unsing the NEAT method. Nothing 
online yet, though.

My next goal is to evolve a snake like object (four bodies connected by 
three hinges) and then something two legged. I'm trying to use Juice to 
design the body - great tool Nate!

> Sorry, I haven't got any details online at the moment, but have a look
> at http://www.bunnyburgers.com/tigger.jpg  for a picture of the beast 
> in action.

That looks incredibly cool! I'd love to see him in action. Let us know 
when you have something to look at!

> I'd previously used the sort of control systems described by Railbert,

Railbert? Never heard of him, nor could I find any online papers - any 
links you could send?

> My reasoning was that the railbert approach might be producing a
> control
> system that was fighting against the natural dynamics, whilst a
> non-prescriptive approach might work in harmony with the dynamics. The
> down-side is that the system is so non-prescriptive that a lot of time
> is wasted simulating completely unfeasible control weightings.

That's evolution for you. I try very hard to kill the bad systems off as 
fast as I can.

> The creature evolves to hops ok if I use both legs in sync, but I've
> only just started on the bipedal running strategies, so it's a bit 
> hit and miss at the moment.

Well, if you want to toss ideas around, I'm game!

cheers,
m