This OpenVR based code can be used to grab positions from the HTC Vive, allow you to calibrate it, and to run a server that broadcasts the positions over the network. You can also visualize data save in .bin file for analysis purpose.

You need to install steam and SteamVR.

Note: There is possibilities that it doesn't create udev rules properly, if the file /lib/udev/rules.d/60-HTC-Vive-perms.rules is not created by the process, you can use the one from misc/.

In this repository python3.* must be used.

You need to install the following dependencies:

pip install -r requirements.txt

Go to Steam -> SteamVR -> Properties -> Betas. And choose beta - SteamVR Beta Update

Add the following options to ~/.steam/steam/config/steamvr.settings:

"steamvr"{
    "requireHmd" : false,
    "forcedDriver": null,
    "activateMultipleDrivers": true
},
"driver_null" : {
    "enable" : true
}

Intel graphics are not supported by SteamVRV. Error (306) : install mesa-vulkan-drivers (needed ?) Error (307) : you can track with this error

After lunching steam, you might need to kill the process vrcompositor.

First, you can edit field_points.json to set up your ground truth positions. Then, run vive_field_calibration.py. You need to have a paired controller then, and to go to each position one by one (indicated in the 3D viewer by an arrow) to tag them.

The field_points.json should contain at least 3 points, and formatted as following:

[
    [x1, y1, z1],
    [x2, y2, z2],
    [x3, y3, z3],
    ...
    [xn, yn, zn]
]

Alternatively, you can pass a file path to vive_field_calibration.py with -p my_field_points.json.

If a consistency error occurs, the joystick will vibrate, and you should start the calibration over.

In the new calibration version, a new file field_points_trackers.json is created with each tracker used to calibrate the field.

{
  "LHR-42A22618": [-2.5, -2, 0],
  "LHR-815E4573": [1, 0, 0],
  "LHR-52015056": [-2.5, 2, 0],
  ...,
  "<trackers_id>": [xn, yn, zn]
}

It's recommended not to place trackers on point of interest.

You can run the viewer using vive_bullet.py script. This will display the trackers and the controllers positions.

The script vive_server_auto_calib.py (old version : vive_server.py) is a server that broadcasts the positions through the network using UDP and protobuf definition from proto/vive.proto.

Note: to avoid spamming, we broadcast UDP messages to 192.168.0.255 by default, this can be changed using -b flag, you can pass <broadcast> to send the packets to all possible addresses

To check, you can also run the vive_bullet_client.py that listens to the network instead of using directly the OpenVR API.

After you killed a server. A binary file with all data from all tracker, base station, and controller are saved in two different file. The first one contain data uncalibrated and the second with the calibrated ones.

You can use generate_protobuf.sh to regenerate protobuf files

OpenVR output seems to be different with what is explained in the official guidelines. Thus, we referred to the vr_tracker_vive_1_0.stl file. The frame we provide is modified so that:

• the origin point is at the bottom of the sensor, centered on the screw hole
• z is upward
• x is facing the board LED

When you closed vive_server_auto_calib.py, two files are generated. The first one is the uncalibrated version and the other one (with _auto_calib at the end) is the file with all data calibrated. To visualize data from a .bin file, you can launch

    python vive_data_visualization.py -l logs/<your log>.bin

Example :

    python vive_data_visualization.py -l logs/2022_09_27-11h32m48s_vive_auto_calib.bin

or

    python vive_data_visualization.py -l logs/2022_09_27-11h32m48s_vive.bin