Various questions regarding RTAI/Simulink/RTW/Comedi

Roberto Bucher roberto.bucher at supsi.ch
Mon Jan 23 19:02:04 CET 2006


Hi 

I'm one of the developper of RTAI-Lab, and I try to answer to some of your 
question.

On Monday 23 January 2006 16:50, janne asdf wrote:
> This entry contains various ("random") questions regarding the above
> mentioned topics, feel free to
> contribute to, or skip, any of them.
>
> I'm a trainee at a company that wants to investigate the possibility of
> using Real-Time Linux for various
> control systems. More specifically, the move from VxWorks to Linux.
>
> I'm not into Control Systems (and Matlab/Simulink etc) myself, which is why
> I've been reading various articles
> to get a grasp on how it all is connected. And basically what I need to do
> in the end is just a to create a
> simple control loop with the Matlab/Simulink chain using a I/O-card that is
> hopefully supported by Comedi.
> All in all, diving into RTAI, learning basic control system theory,
> Matlab/Simulink, etc at once feels slightly
> overwhelming. But I'm here to learn, and expect to gain a lot of valuable
> experience.
>
> So here comes a few questions & thoughts:
>
> Some questions about Simulink etc.
> * After creating a model in Simulink, generating/compiling the code, and
> running it on RTAI how exactly
> (with what?) can the model be monitored/controlled in real time? (gauges,
> displaying values etc). I think
> I've read something about RTAI-Lab? How do you "connect" the running
> control loop with a tool like this?
> Also, I know I can do this on the same machine, but if I want to use a
> WinXP machine for that, can I use
> a RTAI-Lab equivalent seamlessly (Matlab/Simulink/RTW)?
>    This all is a bit confusing still..

RTAI-Lab is born mainly as "substitution" of the "external mode" provided by 
Matlab/Simulink/RTW. The idea is to be free from any CACSD environment, one 
time the RT code has been generated. The "external mode" requires that the 
user buy a full Matlab/Simulink/RTW licence to run and monitor the RT task. 
In our solution, the user doesn't need Matlab/Simulink/RTW anymore, one time 
the code has been generated. You can distribute and monitor the RT task 
without the need of the Matlab/RTW licence! In addition, this method allowed 
to implement a Rapid Control Prototyping environment for the open source and 
free downloadable Scilab/Scicos system.

The "xrtailab" application can be started on the same Linux machine where the 
RT task runs, or on another Linux RTAI machine in the LAN. The possibility to 
monitor a RTAI RT task from a "not RTAI" OS (including WinXP) is given by the 
RTAI-XML project

http://artist.dsi.unifi.it/rtaixml/

> About Comedi:
> I read somewhere on the mailinglist once about the performance of the
> I/O-cards using Comedi, and
> that someone was not satisfied with the performance (sampling rate). And a
> short comment as answer
> in the lines of "because it's not DSPACE. Could someone tell me
> what "DSPACE" is about in this case,
> and what the actual issue was.
The only info about DSPACE is that it is a system connected to 
Matlab/Simulink/RTW, mainly used in the automotive domain. I don't know 
exactly what kind of processorruns in the system (I means a FP DSP but...). 
The system has great performances and an adequate cost.
With the RTAI-Lab sytem we can work with sampling up to 10KHz without 
problems. These are the typical sampling frequencies in my laboratory.

>
> Also, the "RTAI-Lab tutorial - Scilab, Comedi, and real-time control"
> states that DRI support is not
> compatible with RTAI-ADEOS in 2.6.x kernel. Is this still valid? And if
> yes, can someone explain
> what the reason is?
This part of the tutorial has been written by Simone Mannori (
smannori at f2n.it). You can ask him directly about DRI support.

>
> Sorry if some of the questions seem blurry or just "off", it is just I have
> been forced to take in so much
> information at once (not only by diving into RTAI itself), that I do need
> some guidance.
> All answers, comments, links etc. are highly appreciated.
>
> regards,
> Camino

Regards

Roberto

-- 
University of Applied Sciences of Southern Switzerland
Dept. Innovative Technologies
CH-6928 Lugano-Manno
http://www.dti.supsi.ch/~bucher




More information about the Rtai mailing list