Page 1 of 1

Problems with SSVEP mind shooter scenario

Posted: Thu Oct 22, 2020 11:09 pm
by Jose7198
Hi everyone!

I have been trying to run SSVEP mind shooter example scenario and failed in third step (ms-3-CSP-training.xml).

I have tried everything, different versions of OpenVibe among different OS and in every try the execution last excatly 270 seconds and the designer sudenly closes.

Logging a little I found that the problem is in Regularized CSP Tariner box, in Windows it closes the designer with no further details but in Ubuntu I obtained this message:

Code: Select all

[  INF  ] At time 270.211 sec <Box algorithm::(0x0000228f, 0x00004370) aka Regularized CSP Trainer> Received train stimulation - be patient
[  INF  ] At time 270.211 sec <Box algorithm::(0x0000228f, 0x00004370) aka Regularized CSP Trainer> Data covariance dims are [66x66]. Number of samples per condition : 
[  INF  ] At time 270.211 sec <Box algorithm::(0x0000228f, 0x00004370) aka Regularized CSP Trainer>   cond 1 = 5 chunks, sized 1500 -> 7500 samples
[  INF  ] At time 270.211 sec <Box algorithm::(0x0000228f, 0x00004370) aka Regularized CSP Trainer>   cond 2 = 5 chunks, sized 1500 -> 7500 samples
[  INF  ] At time 270.211 sec <Box algorithm::(0x0000228f, 0x00004370) aka Regularized CSP Trainer>   cond 3 = 5 chunks, sized 1500 -> 7500 samples
[  INF  ] At time 270.211 sec <Box algorithm::(0x0000228f, 0x00004370) aka Regularized CSP Trainer>   cond 4 = 5 chunks, sized 1500 -> 7500 samples
[  INF  ] At time 270.211 sec <Box algorithm::(0x0000228f, 0x00004370) aka Regularized CSP Trainer> Compute CSP Begin
./openvibe-designer.sh: line 58: 20438 Segmentation fault      (core dumped) /home/johanna/Documentos/Titulacion/OpenViBE/meta/dist/extras-Release/bin/openvibe-designer
Any idea what could be the problem? I am using an OpenBCI board and used the electrode setup detailed in the documentation (Oz, CPz, Pz, Iz, O1 and O2)

Re: Problems with SSVEP mind shooter scenario

Posted: Fri Oct 23, 2020 8:03 am
by Thomas
Hi,

Thank you for the bug report, we will look into it. Thanks for starting the investigation and pointing at the failing box.

Cheers,
Thomas