Subject:
|
Re: controlling RCX from PC?!
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Wed, 27 Oct 1999 19:31:01 GMT
|
Original-From:
|
nik heger <nik@=spamcake=stagecast.com>
|
Viewed:
|
2058 times
|
| |
|
|
>
> Remember the NASA Pathfinder? Signal round-trip was about 30 minutes! Still,
> they did some interesting research. Try to put more decision-making
> capabilities in the RCX itself and let the PC handle its behaviour on a more
> global level.
> What I'm trying to say is: Let the RCX do the driving and the PC do the
> navigating. The 'driver' can decide when to stop for an obstacle. It can
> also ask the navigator which way to go at an intersection. This way there
> will be a lot less messaging going on, you can speed up your robot and still
> avoid accidents. You will probably have to redesign your software to achieve
> this.
thanks matthew and arjen!
that's exactly what i am going to do.
what i did not realize at first is that 2400 baud actually *is* way
too slow, as 2400 bits/s is only 240 bytes per second. a message has
a length of around 10 bytes (depending on the op code) and the
response is at least 5 bytes greater than that, so it takes about 25
bytes to send a command and receive a response. that is about 9
communications attempts per second in the best case (no lost packets,
no switch time for send/receive). 3 of these is requesting the status
of the sensors. too slow for real time control (at least the type i
am doing).
thanks again for your help.
nik
ps: NASA probably had more resources than i do. while i don't know
their numbers, i am sure it was more than one person and more than 2
days :-)
______________________________________________________________________
nikolaus heger
nik@stagecast.com PGP fingerprint:
650.354.0735 970A 0D39 41BD A660 AACB 674B 747A
79C0 55B8 8F9D
______________________________________________________________________
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
