|
[DPRG] Help with Elementary School Project
Subject: [DPRG] Help with Elementary School Project
From: Dale Wheat
dale at dalewheat.com
Date: Sun Apr 22 22:34:53 CDT 2007
Winefred,
I'll have a go at the first one:
> 1. How do autonomous robots collect information and make decisions off of it?
Most autonomous robots that I am familiar with use electronic computers to
control their behavior. These computers run programs that are written by humans
and attempt to anticipate the events and surroundings that the robot might
encounter. The robots usually have several different kinds of sensors that give
them information about their environment. Some examples include bumper switches
that can tell when the robot has run into something, light sensors that can tell
the difference between light & darkness, microphones that hear sounds, and even
more specialized sensors. These sensors correspond to very primitive versions
of our human senses.
Depending on how complex the computer programs are that are controlling the
behavior of the robot, a sensor input might cause one or more reactions. For
example, if a small robot runs into the wall, a bumper switch might could
indicate this, and the computer program, which is anticipating this very
happening could say, in effect, "Hey! I've bumped into something, so I'd better
stop, back up and go in a different direction now." Now if this same robot is
supposed to be a bulldozer, and is supposed to push things around, it might
think "Hey! I've bumped into something... I better just keep on pushing in this
same direction, because that's what I want to do."
Because computers have memory, a robot that has a computer brain can have
memory, too. If a robot keeps bumping into things over and over again, it might
remember each time and after "too many times", it would decide that it was stuck
and quit trying to get away. How many times is "too many times" depends on what
the original computer programmer decided when the program was written. I have a
robot that has an emotional state. It drives around amilessly and the longer it
goes without bumping into anything, the happier it gets. When it gets happy, it
drives faster and it makes a happier kind of noise. It also has a red and a
green light, and if it is mad the red light comes on, and when it is happy the
green light gets brighter. If it bumps into an obstacle, it gets mad, stops,
and makes a mad & irritated kind of noise. In a very short amount of time, it
starts to get over it and begins to get happier again, or at least until the
next time it bumps into something. Note that the robot very quickly "forgets"
that it was mad, and always reacts the same way every time. This is a simple
example of a robot having a "memory" about itself and what it is doing. In
reality, it's not an actual emotion, but simply a number that goes up and down
depending on what happens to the robot.
So the decisions a robot makes depend on what kind of information it already
has, some of which may be incorporated into its behavioral program, as well as
information that it gathers about its environment from its sensors. The
decision-making process is really just a mathematical formula that describes the
possible outcomes in terms that the robot can understand and act upon.
Thanks,
Dale Wheat
(972) 486-1317
(877) DALE WHEAT
http://dalewheat.com
More information about the DPRG mailing list
|
