[ODE] Hi everybody

[travis cline]

Hey, just signed up for the list here.

I'm a young coder just recently involved in neural nets and artificial life 
(and as of a few days ago, ODE)

Although I'm fuzzy on a few concepts still ODE has come easily, it performs 
very well (surprisingly, not the best graphics card here).

I'm currently putting together an artificial life program similar to Gene's 
( www.erachampion.com/ai )

I'm going to have many instances of 3 wheeled vehicle (its similarity to the 
test_buggys vehicle is purely coincidence =P ) learning to pick up rewards 
(neural net and GA).

I have everything ready to go, my net is set up and I have X many vehicles 
and rewards going into my environment and wrapping around in a set area.

I was about to write up the net inputs when I decided I wanted to change how 
I represented its environment.  I had it set up where vectors to the closest 
reward was fed into the net.  I will eventually add inputs to help them 
avoid each-other (I'm not allowing contact joints between vehicles as of 
now) and I think an 'eye' system will be better suited.

My overall question here is.. will it work?
Can I send out 10 or so invisible boxes to serve as eyes to feed back 
collision information and keep it running smoothly?  And it seems offhand I 
wouldn't be able to give info on the distances of the objects..

I will implement my radarlike system now but I feel it nessicary to change 
soon.

Thank you
Travis Cline

_________________________________________________________________
Protect your PC - get McAfee.com VirusScan Online 
http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963