Overview

Receiver

There are different receiver classes that provide ViconFrame data from different sources:

• ViconReceiver: Uses the Vicon Datastream SDK to connect to a compatible Vicon application (e.g. Vicon Tracker).

• PlaybackReceiver: Loads a previously recorded Vicon frames from a file and plays them back.

• JsonReceiver: Loads a single frame from a JSON file and always provides the data of that frame. This is only meant for testing purposes.

ViconTransformer

ViconTransformer

Gets Vicon frames from a receiver and transforms subject poses such that they are relative to a given “origin subject” (instead of relative to Vicon’s origin).

Basic example in C++:

#include <vicon_transformer/vicon_receiver.hpp>

using vicon_transformer;

int main()
{
    // config is initialised with default values
    ViconReceiverConfig config;

    auto receiver = std::make_unique<ViconReceiver>(
        "vicon_pc_hostname", config);
    receiver->connect();  // connect to Vicon server

    ViconTransformer vt(receiver,  "my_origin_subject");

    // get new frame from receiver
    vt.update();

    if (vt.is_visible("robot_arm_marker"))
    {
        Transformation robot_tf = vt.get_transform("robot_arm_marker");
        // ...
    }
}

Basic example in Python:

from vicon_transformer import ViconReceiverConfig, ViconReceiver, ViconTransformer

def main():
    config = ViconReceiverConfig()

    # instead of using a context manager, you can also call `receiver.connect()`
    with ViconReceiver("vicon_pc_hostname", config) as receiver:
        vt = ViconTransformer(receiver, "my_origin_subject")

        # get new frame from receiver
        vt.update()

        if vt.is_visible("robot_arm_marker"):
            robot_tf = vt.get_transform("robot_arm_marker")
            ...

o80 Driver/Standalone

This package provides templated driver and standalone classes to integrate ViconTransformer through o80:

• o80Driver

• o80Standalone

Since o80 requires observation data structures to be of fixed size, the dynamic ViconFrame class can unfortunately not used as observation type here. Instead, FixedSizeViconFrame is used. This has two consequences:

1. The number of subjects has to be known at compile time.

2. Subject names are not included in the data structure. Instead, their poses are given in an array in an order that has to be specified at compile time.

This is done through the template arguments of o80Driver (see there for more information).

For an example, how this is used in practise, see the implementation in pam_vicon.

Executables and Scripts

In all cases, you can run the executable with --help to get a complete list of options.

vicon_record

Record data from a running Vicon system and safe to a file. The file can then be played back using the PlaybackReceiver.

vicon_record <hostname or IP> output_file.dat -d <duration in seconds>

vicon_print_data

Print data from either a running Vicon system or a previously recorded file:

vicon_print_data <host or file>

Important

vicon_print_data prints the poses as reported by the Vicon system, i.e. it does not transform with respect to some origin subject!

vicon_print_data_py

Python-version of vicon_print_data (see above). It doesn’t support playback of recorded files, apart from that it is mostly equivalent to the C++-version and mainly serves as an example on how to use the Python bindings.