Subject:
|
RE: Ideas for a RIS 2.0
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Tue, 17 Aug 1999 15:19:18 GMT
|
Viewed:
|
679 times
|
| |
|
|
Ok, I just have to jump in here. This thread is a really good idea, and I
hope that TLG monitors it...but I get the feeling that I'm hearing folks
bemoan why they can't build something because of shortcomings in the RCX, and
that's like not building because you need that "special" part from the 8880...
The replacement firmware out there (LegOS and pbFORTH) adress almost every
item below. Normally I don't copy so much text.
> More I/O ports would be handy. But a little ingenuity can expand on what is
> already provided.
Not much we can do about that with firmware, except make it easier to talk to
other RCXs - see below.
> What cannot be expanded is the memory capacity of the RCX, or its processong
> flexibility. The idea of simply using the RCX as an interface between the I/O
> ports and a remote computer is not really an acceptable situation for anyone
> interested in building autonomous devices.
>
> So ..
>
> it needs memory and enhanced processing options to provide the ability to
> handle
> 2D arrays for mapping and maze processing.
Replacement firmware has this already. The pbFORTH image is about 13K, and leaves
about 15K free for more code, variables, arrays, etc. You have to learn a new
language, but if you think of it like learning TclTk or Perl and use enough
to get the job done, then it's not so bad. :-)
> it needs extended processing power to handle route planning and collision
> avoidance.
>
> it needs numerical processing extensions for 32 bit processing. (16 bit
> intermediate
> results overflow far too easily!)
pbFORTH offers double wide 32 bit integers and basic math. It's pretty easy to go through
a numerical algorithms book and cobble up fixed point trig as needed. Often, the 16 bit
math is good enough, the FORTH math often holds 32bit intermediate results so that
numbers dont overflow. The */ operator which multiplies 2 16bit numbers and then divides
the 32bit result by a 16bit number is a good example. It's great for scaling numbers.
> and of course, a much higher speed link to the host PC for rapid download and
> debugging of what would now be much larger fragments of code. RF would be my
> choice, but IR is good. A temporary tether for program development using a
> really
> high speed serial link would be ideal.
Yes, that would be nice. pbFORTH offers arbitrary ASCII strings to be sent to/from
the RCX. In fact, the TckTk console I wrote lets you communicate in English with
your RCX - no complicated bytecodes to learn. I'm thinking of trying to increase the
baud rate options for the next release.
The O'Reilly book coming out this fall will explain all of the "alternative"
programming environments (NQC,LegOS,and pbFORTH) in great detail. I think that
once the community gets their arms around these tools, we'll see some pretty
amazing things out of the little yellow box...
Cheers,
Ralph Hempel - P.Eng
--------------------------------------------------------
Check out pbFORTH for LEGO Mindstorms at:
<http://www.hempeldesigngroup.com/lego/pbFORTH>
--------------------------------------------------------
Reply to: rhempel at bmts dot com
--------------------------------------------------------
|
|
Message is in Reply To:
 | | Ideas for a RIS 2.0
|
| More I/O ports would be handy. But a little ingenuity can expand on what is already provided. What cannot be expanded is the memory capacity of the RCX, or its processong flexibility. The idea of simply using the RCX as an interface between the I/O (...) (25 years ago, 17-Aug-99, to lugnet.robotics)
|
4 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
