VRPN

[Lorena]

Hello all,

I am working with VRPN now and I was wondering how should I use the 'Button VRPN Server' box. I want to send to an external application 3 different classified status: left or right; blinking; hand movement. I think I should use one Button VRPN box with 4 inputs. Are these inputs to the VRPN box the classifiers' outputs? I think the Button VRPN Server box configuration should be the one shown in the attached image, is that configuration correct?

I'm not sure about this appoach, so I have some questions.
* Are the classifiers' outputs just labels saying 'left' or 'right'? or are they values that I must interpret?
* For the classifier that detects whether the movement is to the left or to the right, can I use two different buttons (one for the left label and one for the right label)?
* What labels should I use for the button off states? Do these labels come from the classifier as well?

I know some of my questions are very basic, but I want to understand clearly how this VRPN comunication works.

Thanks for any help!

Lorena

[uahmed]

Hi Lorena,
Yes, the output of the classifier will actually act as the input to the VRPN server.

The output of the classifier are not just labels, they are actually values, that is how the process actually goes. Left is different from right, and so on.

I guess you can use to different buttons as well. For the off states, I will recommend that you may use the Hit and Trial technique, and find out how you meet your objective in the best way. Hope this helps.

Thanks.

[Lorena]

Hi Uahmed,

How do you recommend me to set my parameters for the VRPN server?

This is what I want to perform: I want to send three different commands to the external application: 1) hand movements to the left or to the right; 2) eye blinks; 3) tongue movement.

I use three separated pipelines with one classifier each. The first classifier decides if there is hand movement to the left or to the right; the second one decides if there is an eye blink or not; and the third one decides if there is tongue movement or not.

I think the settings I show in the figure are wrong, because there is one classifier to decide left or right hand movement, and there is no 'NonTarget' label, so that configuration wouldn´t work, right?

Thanks a lot for your help!

Lorena

[uahmed]

Lorena,
Did you see the tutorial? If not, then here is the link.

http://openvibe.inria.fr/vrpn-tutorial- ... plication/

I am also trying to perform the same task as of yours, and it is my FYP. But the problem I am facing is that I have not been able to use the VRPN server successfully, since am weak in programming. If you can help me to get through that stage, I am sure I can help you in accomplishment of the task. The use of on and off buttons wont be an issue then, once I get through it.

Thanks.

[Lorena]

Hi Uahmed,

What kind of problems are you having with the VRPN Server? Are you using analog or button VRPN? Of course I will help you if I can!

Lorena

[uahmed]

Rite now, I am just working over the example of vrpn given in the tutorial. I guess I have not been able to successfully use the vrpn server, may be there are some errors in the compilation of VRPN SERVER.

I want to know that how can I link C++ and VRPN Server. From that tutorial of keyboard and sinus oscillatior, I copied that main code, added the two header files, and then tried to run, but in vain. I got no output on C++. Which files do I need to add more? Do I need to use the VRPN GENERIC as well? Please guide me, since I am not much deep into the programming stuff.

Thanks.

[Lorena]

Well Uahmed, I am not working with the VRPN implementation yet, just preparing the scenario and the info to send to the external application. So, when I get to that part I will tell you how it worked for me (I am not so experienced in programming, so I will see what I can do).

Lorena

[uahmed]

Okay Lorena, wish you best of luck. I guess our goal is same, so we can discuss our techniques and steps implemented so far to reach the current stage (like classifier training etc), it may help us to improve our algorithm. Let me know if you feel it as good idea.

Thanks

[uahmed]

Hi Lorena,
I have almost accomplished the task. We can discuss it some time.
Thanks

[Lorena]

Hi Uahmed,

It is very good news that you are reaching your objective, of course we can discuss our experiences, I think it would be very helpful for both of us

Lorena

[uahmed]

Hi,
Thanks I guess I can help you out now, we can talk at skype. Or whatever suits both of us more.

Thanks

[uahmed]

Hi Lorena,
I am getting one problem in getting the stimulations. I keep getting the stimulation continuously, I checked that via stimulation listener too.

Thanks

[Lorena]

Hi Uahmed,

What stimulations do you want to send? From what you say, I guess you want to send one stimulation and stop sending, is that correct? So the problem is that you are sending it every x seconds all the time?

Lorena

[uahmed]

Yes. Like From the time I run my online scenario, I keep getting the stimulations, even when i am completely neutral. My online signals just don`t seem to have any effect.
Thanks

[uahmed]

Lorena,
During the training session, I had got 98% accuracy. But, I had fast forwarded the session... Is this the issue?
Thanks.