Subject:
|
Re: Bump switches and "aggression"
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Wed, 3 Dec 2003 17:49:23 GMT
|
Original-From:
|
Jim Choate <ravage@einstein.ssz.com+StopSpammers+>
|
Viewed:
|
1644 times
|
| |
|
|
On Wed, 3 Dec 2003, Steve Hassenplug wrote:
> How exactly, do you want to trigger this 'fight or flight'?
Easy the bump switch jumps into a RNG, and it's important to understand
that the output of a RNG does -not- need to be equi-probable, and it
decides. Watch animals when they're startled, they don't stop and think,
they jump around for a second and then take off.
> It seems to me that a rabbit (for example) would use both of the methods
> above. If it's running along, and it's nose smacks into a tree (I mean bumps
> into a tree) the behavior is very deterministic.
Here's the problem with your approach, blind rabbits don't live long
enough to run into trees. What a real rabbit does is either freezes or
takes off in a pretty much random direction, if they happen to hit
something (say a pet rabbit in a bedroom that's startled with the phone
rings) like a wall they reverse their trajectory and begin their avoidance
strategy again. They continue this until they get under cover or are
caught.
How does the rabbit determine cover? It's eyes. How does a robot determine
cover? Changes in light intensity via optical sensors.
Now, a rabbit obviously has a more complicated behavior pattern than this.
Why?
Because it has higher levels of processing going on that gate or squelch
these behaviours. An example is what a rabbit does once it's a bit away
from whatever startled it, it keeps going in -roughly- the same direction.
What does that mean? It means that the process that causes the random path
changes is still going on. How might one model that? I'm working on
another project for this. Consider a ball. It has a weight in it that
makes one end come to the bottom. Around that weight there are effectors
that will come out at random times and push the ball in some direction.
How can we modify this with a 'cover' behavior? Let's assume we have three
effectors, we need three light sensors. Do we want approach or avoidance
to light? Most cover is dark, it blocks sight and therefore light. So, the
RNG in the ball-bot (I call it "Rho") keeps on trying to trigger the
effectors, but we gate them with a AND function that is based upon light
intensity. We do a comparison of the three sensors, which ever is lowest
tells the others to turn off their effectors. When the RNG next gets
around to hitting the effector for the dark sensor (I assume the sensors
are opposite their effectors) the ball rolls toward the dark.
> But, if it detects ground vibration, or possibly a noise from an unknown
> direction, that is when the 'fight or flight' kicks in.
Actually not, I've hunted and had pet Rabbits for more than 40 years. What
a Rabbit does do is run off directly away from the vibration if there is
time. If not the Rabbit freezes. They're behavior is very predictible.
It's what makes them easy to hunt.
I can't count the number of times I've been bopping along in a pasture
with other hunters spread out in a line and actually kicked rabbits. Sometimes
they run, sometimes they don't they just sit there and let you kick'em
again. You can litteraly reach down and pick them up, if you don't mind
the chance of having a buzz saw in your hand. Possums do this sort of
stuff too.
Thanks for the feedback.
-- --
Open Forge, LLC 24/365 Onsite Support for PCs, Networks, & Game Consoles
512-695-4126 (Austin, Tx.) help@open-forge.com irc.open-forge.com
Hangar 18 Open Source Distributed Computing Using Plan 9 & Linux
512-451-7087 http://open-forge.org/hangar18 irc.open-forge.org
James Choate 512-451-7087 ravage@ssz.com jchoate@open-forge.com
|
|
Message is in Reply To:
 | | Re: Bump switches and "aggression"
|
| (...) How exactly, do you want to trigger this 'fight or flight'? It seems to me that a rabbit (for example) would use both of the methods above. If it's running along, and it's nose smacks into a tree (I mean bumps into a tree) the behavior is very (...) (21 years ago, 3-Dec-03, to lugnet.robotics)
|
60 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
