Subject:
|
Cybermaster for show-me training
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 9 Sep 1999 20:00:41 GMT
|
Viewed:
|
445 times
|
| |
|
|
In lugnet.robotics, lego-robotics@crynwr.com (Laurentino Martins) writes:
> [...]I happen to have a CyberMaster and I noticed all you have said
> about the Show-me Training (R)
No "(R)" please! "show-me" is registered six times as a trademark (display
stands, restaurant services, spot lights, AV equipment rental, baseball
exhibition, and barbeque sauce -- really!) so let's not get any of them upset
(-:
> applies perfectly to the CyberMaster and much better than the RCX
> since no extra inputs need to be used for that purpose.
Does the CyberMaster allow reading the tachometers even when the motors are in
the "float" off state, or only when they are on?
> [...] There's only one catch, you don't know the direction they are
> rotating. But then again, neither the RCX
Not true -- with an appropriate resistor network the zero-point can be biased
such that one direction of rotation produces a positive deviation and the other
direction produces a negative deviation. You have to do this anyway, because
the sensor input voltage has to be non-negative and the (hypothetical) driven
motor is capable of generating a negative voltage. The same approach would
probably work on Cybermaster.
> Also, it seems the algorithms are too heavy to be processed the RCX
> itself [...]
Well, there's the main reason I don't want to switch to Cybermaster. I actually
want my robot to be completely autonomous, so I can retrain it at a friend's
house without having to install software on his computer. In fact, I'm hoping
to bring it to a sci-fi convention next month...
But I don't want to discredit this approach on the Cybermaster. You'd still
have to hook up the motors through a resistor network to read the polarity
(direction of motion), but the tachometers would give a more accurate speed
reading. And, the use of the computer makes it easy to do nonlinear or
3-or-more variable regression analysis, and lots of more complicated
mathematical models too.
> so there's nothing better than a program remotely monitoring the
> CyberMaster (using the Radio link, remember?) and generating a NQC
> program of the behavior performed. That program could be downloaded
> with a key press.
Now there's an interesting idea. That would make it so you need the computer
only for retraining. That's a pretty interesting idea. Of course, if you want
to use NQC you'd have to limit yourself to the simpler mathematical models, but
that's beside the point.
> It seems so easy that's weird nobody ever remembered trying it with
> a CyberMaster!
> Any volunteers? :-)
I think the general idea of recording motor movements, then somehow translating
it into a program that gets downloaded back to the robot is a splendid idea
ripe with possibility. And it does seem that the Cybermaster is a natural for
this.
- Robert Munafo http://www.mrob.com/
LEGO: TC+++(8480) SW++ #+ S-- LS++ Hsp M+ A@ LM++ YB64m IC13
|
|
Message has 1 Reply:
Message is in Reply To:
 | | Re: show-me training
|
| I know that the RCX is the state of the art LEGO computer, but I happen to have a CyberMaster and I noticed all you have said about the Show-me Training (R) applies perfectly to the CyberMaster and much better than the RCX since no extra inputs need (...) (25 years ago, 9-Sep-99, to lugnet.robotics)
|
17 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
|
|
|
|
