Subject:
|
Re: RCX simulator ?
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Mon, 16 Sep 2002 15:13:08 GMT
|
Viewed:
|
1334 times
|
| |
|
|
Hi all!
> Simulating a RCX can mean many things:
> 1. simulating the software and viewing internal and external "signals"
> 2. simulating the IO-ports, so you can also test homemade sensors/actuators
> 3. simulating simple movements with a left/right motor or power/steering
> motor
> 4. simulating bumping against walls (just binary bump or not bump)
> 5. simulating IR transmission between RCX and Tower or another RCX
> 6. simulating the real torque of motors (which is indeed impossible,
> never seen 2 motors that where exactly the same)
> 7. simulating real mechanics, you're pointing to
> 8. simulating the heath development of motors of motors etc...
> 9. simulating the degradation of the batteries
> 10. ... many other things
We don't need to simulate all of the mechanics because we've got the real
thing. What about remote controlling the complete robot with the brain
inside your PC? In the case of simulation we could degrade the RCX to a
stupid IO device. Take its sensor's input and set its output i.e. its motors.
The RCX would need a very simple program (written in a few minutes ;-) )
- to respond to IR messages,
- to send it's sensor values (only if they have changed),
- to set the type of sensor (light, switch etc.),
- to control the motors
An efficient communication protocol could handle this even with 4800 baud in
a sufficient short time.
The PC on the other side relies on the sensor inputs and controls the motor
outputs, as if it were the RCX. Then the PC could simulate NQC (done within
a few hours) or even simpler brickOS (done within a few seconds (ok, not by
me)).
For brickOS you would only need to write some IO libraries (that map brickOS
commands to communication protocol commands) and some wrappers (e.g. for
threads etc.). You could then compile and debug your programs as usual on
your PC, but with great new debugging abilities in respect to the normal
development procedure on the RCX.
All other outputs like the display and the sound could be handled by the PC.
The only drawback I see so far is that of
- time delays (sensor reading -> reaction. Fix: put some cushions under the
table of which your robot tries to detect the edges)
- failures in IR communication because of obstacles etc.
Unfortunately I'm currently finishing the interior of my new house and my
wife thinks this is more important than building robots. So, is anyone
interested in just hacking down a few lines of code? ;-)
Greetings
Thomas
|
|
Message is in Reply To:
 | | Re: RCX simulator ?
|
| (...) I cann't disagree with you, ... but I live in an surrounding where even 99% is using this M$-OS. And I'm just too simple to get Linux working ;-( I'm not a software programmer, and learning one computer language was difficult enough for me. (...) (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
|
|
|
|
