Wrapper for VICON DataStreamSDK v1.12 for network online streams via VICON.
Use contents of ViconDataStreamSDK_1.12.0.145507h__x64.exe to install the SDK on target computer. Ultimately, code in this repository makes use of the DotNET support for ViconDataStreamSDK, as the C API was confusing for MATLAB to parse.
- Connect via orange ethernet cable, using static IP address
192.168.10.50with gateway192.168.10.1. The192.168.10.1will go in place oflocalhostwhen you are running this library not on the VICON Nexus computer. - Use
main.mto make sure you can connect and count/enumerate the available devices. - Check
config.yamland make sure that the configurations are correct. The key one to check isLSL_Folder, which should be wherever you have installed MATLAB LSL on your local computer. It should point to the folder containinglsl_loadlib.m.
Use the script stream_trident_to_lsl.m to stream the IMUs to LSL. Use the script stream_trident_and_markers_to_lsl.m to stream the IMUs + calibrated subject marker set to LSL.
If you are going to try streaming a new marker configuration for the first time, follow these steps:
- Calibrate cameras.
- Place markers.
- Create a new subject for your session. Don't do anything to the subject.
- Go to the "Subject" tab on the right. Capture a session. You will now be in "Offline" mode. You must click through and create Segments for any segment you want to create. If you don't know what that means, probably don't do this.
- After creating segments/associating markers to them, you may wish to change the default naming schema, which is "Segment" for the "N-th" marker (1-indexed) per Segment. You can update the names, but note that these are now the default names that the DataStream labeling template will be spitting out. To get the updated marker names to save as well as any other properties of the new markers, you should now right click your subject and Click "Save as Labeling Template and Attach..." then name it as you want. Now, your labeling template is attached to the current marker set and should have the correct names etc.
Probably don't do this either- it will be faster just to re-fit the model at the beginning due to minor idiosyncracies/inconsistencies in the individual marker rigid relations which will cause the VICON streams to become angry (i.e. the model won't know how to scale its likelihoods properly for the individual markers or however VICON does things "under the hood").