Subject:
|
Re: RCX simulator ?
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Sun, 15 Sep 2002 07:19:47 GMT
|
Original-From:
|
Steve Baker <sjbaker1@airmail.net[AvoidSpam]>
|
Reply-To:
|
SJBAKER1@nospamAIRMAIL.NET
|
Viewed:
|
1335 times
|
| |
|
|
Dave Baum wrote:
> I hoped other
> people would write appropriate front ends (GUIs that might let you
> interact with the RCX's sensors, or perhaps even set up a virtual
> environment for the simulated RCX to move around it, whatever).
That's clearly the next step - the question is - does anyone want
it badly enough to do it?
Personally, I'd rather just run my programs on the real RCX.
It seems to me that RCX programs are generally simple enough that
it's easier to debug them on the RCX than it would be to set up
some kind simulator to behave sufficiently similarly to the real
robot.
Most of the problems I get into on the RCX are things that a simulator
would be *very* hard pressed to get right.
For example, I recently had a robot with bump sensors on both front and
back of the robot (wired in parallel to save sensor inputs) - and code
to make the beast change direction if one of them was hit. However, due
to a programming error, I'd neglected to allow for the fact that the
sensor stayed depressed for *many* iterations of the program after it
had commanded the motors to go into reverse. Also, the bump sensors
didn't go cleanly from ON to OFF, instead reporting something more like:
ON ON ON ON ON OFF ON OFF OFF OFF OFF OFF OFF
^^^^^^
Nasty!
Whenever the robot hit anything it would go into a frenzy of smashing
itself repeatedly into the wall! It wasn't hard to fix - but it was
a little unexpected.
However, this bug probably wouldn't show up in the idealised world of
some kind of simulator unless it was sophisticated enough to figure
out the slop in the gears between the motor and the drive wheels,
the inertia of the robot, the differences between motors, the amount
of depression that has to be applied to a switch before it will trigger,
the change in the voltage going through the sensor when the motors
change direction, the friction between the drive wheels and the carpet...
all manner of complicated things.
I just can't see us ever getting that much sophistication into the
simulator - and until we did, it still might not be much use for
debugging robots. You'd get your program working magnificently in
the simulator and then find they'd work *nothing* like that in the
real world.
> I think an important point is that incremental projects are more likely
> to get off the ground than the "big bang" project where you spend
> hundreds or thousands hours before getting to any meaningful release.
> There's a lot of positive feedback when you can quickly get to something
> that is useable.
Yes. I've worked on several 'big bang' OpenSource projects - and they've
mostly been disasterous (or at lease very, very painful to get finished).
The projects I've worked on that were incremental in nature were (without
exception) vastly more successful - and more enjoyable to work on.
----------------------------- Steve Baker -------------------------------
Mail : <sjbaker1@airmail.net> WorkMail: <sjbaker@link.com>
URLs : http://www.sjbaker.org
http://plib.sf.net http://tuxaqfh.sf.net http://tuxkart.sf.net
http://prettypoly.sf.net http://freeglut.sf.net
http://toobular.sf.net http://lodestone.sf.net
|
|
Message is in Reply To:
 | | Re: RCX simulator ?
|
| (...) I stopped working on it because there wasn't much demand. I never intended to write a full-blown simulation environment. I hoped other people would write appropriate front ends (GUIs that might let you interact with the RCX's sensors, or (...) (22 years ago, 15-Sep-02, to lugnet.robotics)
|
32 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
Active threads in Robotics
|
|
|
|
|
|
