Subject:
|
Re: Bump switches and "aggression"
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Wed, 3 Dec 2003 18:18:09 GMT
|
Viewed:
|
1679 times
|
| |
|
|
In lugnet.robotics, Jim Choate <ravage@einstein.ssz.com> wrote:
> On Wed, 3 Dec 2003, Gordon Elliott wrote:
<snip>
> > Now direclty on bump sensor vs. other senses: If moving in a particular
> > direction, and robot had _directioal_ bump sensors,
>
> There's the fundamental difference there is -no- reason to have
> directional bump sensors. If you are moving along a path and a bump sensor was
> hit that is all you need to know to change your path. You don't need to
> know in which direction. An Ameoba would be a good example.
<snip>
I'm very new to this particular thread, but I like Gordon's
approach--directional sensors are needed. If I'm travelling forward and
something bumps me from behind, I need to know that--If my conditional response
is to do the opposite of what I'm doing when I get hit, then basically I'll back
up right into what bumped me.
I remember having bumpers all around one of my block stacking 'bots' I tied
them all together and programmed 'When bumped back up, then after elapsed time,
turn a bit and go forward again, with the intent that I'd hit a wall, back up,
turn a bit, and continue on.
Well, in the arena one particularly feisty 'bot came and hit mine from the back,
so, natch, my 'bot backs up, right into the 'bot that hit it. That 'bot
proceeded to back up, slower than mine, and mine backed into it again,
triggering the bumper. Well, this went on for a few more hit 'n back-up's, and
we all had a good laugh, and I subsequently modified my code remembering this
very instance.
If my 'bot 'knew' it was getting hit from behind, it would have known to go
forward, or forward/turn, in order to 'get away' from the offending 'bumper'.
Even if you progammed random stimulus, there would still be a chance for my
'bot' to go backwards, right into the 'jaws' of my opponent. If it were me,
random movements *away* from the direction of the bump is the best alternative.
And to do so, you need to know from which direction the bump came.
Anyway, collection vs. debate--this is swinging more towards debate :)
Dave K
|
|
Message has 2 Replies:
 | | Re: Bump switches and "aggression"
|
| (...) go (...) 'bumper'. (...) my (...) me, (...) alternative. (...) There is one solution for staying 100% Lego and putting 4 sensors on the same RCX input. Therefor you should be able to sense which direction you were getting bumped. Lego made (...) (21 years ago, 3-Dec-03, to lugnet.robotics)
| | | Re: Bump switches and "aggression"
|
| (...) Actually you don't, you want a different level of sensors to evaluate the environment (eg ir object sensors). You're just complicating your design by putting too many functions on each layer of sensors. For example, something bumps you from (...) (21 years ago, 13-Dec-03, to lugnet.robotics)
|
Message is in Reply To:
| | Re: Bump switches and "aggression"
|
| (...) You should watch my three dogs hunt. If all you've ever seen is a single dog hunt then you know nothing of how dogs actually hunt. They're pack animals, a single dog can't survive in the wild more than a few weeks. Yet three sub-200lb dogs can (...) (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
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
