DPRG List  

[DPRG] Robots in te news

Subject: [DPRG] Robots in te news
From: Blake Miller blakeage at hotmail.com
Date: Fri Nov 2 22:01:45 CST 2001

I'm working on a video game at school (it's a side scrolling Martial Arts 
fighting game).  I chose to do the AI for the game because I'm fascinated 
with how one can re-create the thought process of a human in a fight.

For instance, my characters (the enemies in the game), if programmed with 
lazy intelligence, would do something like move around randomly on the 
screen, dodging the main character.  What is more interesting; however, is 
creating a character that follows a thought process *similar* to a human 
martial artist.

For instance, how should the enemy react to the main character throwing a 
punch?  Well, in a real battle, it would depend on several things - the 
range of the punch that was thrown, what moves were executed in the past 
following this move, where in space the fist is, the air speed velocity of a 
swallow, etc., etc., etc.

Sure, I can't and probably never will be able to consider every situation 
that occurs in nature, but I can get close.

What would be really cool is if I could port the logic to an artificially 
intelligent battling robot : ), and we could have smart wars : )


----Original Message Follows----
>From: Chuck McManis <cmcmanis at mcmanis.com>
To: dprglist at dprg.org
Subject: Re: [DPRG] Robots in te news
Date: Fri, 02 Nov 2001 13:44:51 -0800

At 01:31 PM 11/2/01, Mike McCarty wrote:
>Since no one knows how people think, I believe you have set yourself an
>unachievable goal.

That would be true only if you also believe that no one will _ever_ know
how people think.

It is essential that every roboticist have an "unachievable" goal to
prevent them from achieving early success and losing interest. Given such a
goal it becomes useful to break off "sub goals" that are more achievable.
For example, build a platform that can balance on two feet. Or build a
platform that can get from point A to point B and back again reliably.

My goal, established over 15 years ago was to build a robot that would, on
voice command, come out from its recharging station, fetch me a cold Diet
Dr. Pepper (replenishing the supply in the refrigerator if necessary),
return to me with the beverage, wait for me to take it, and then return to
its recharging hutch.

In 1984 this goal was "unachievable" because it was impossible to put
enough compute power on the robot to navigate and drive manipulators and do
vision processing. Today that can be handled by last years laptop.

In 1984 a set of Lead-acid batteries that could power the computer that
could navigate the top floor of my house weighed about 52 lbs (2 x 12V car
batteries) now the laptop carries just over two pounds of Li-ion batteries.

So the goal of building a thinking humanoid robot is a good one, it is
currently difficult to do although there are working examples of the
mechanics being possible. Given the rate of increase in mips/pound I think
the software will be doable in 10 to 15 years as well.


DPRGlist mailing list
DPRGlist at dprg.org

Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

More information about the DPRG mailing list