Skip to main content

Frames

The page on TF in ROS discusses the importance of spatial transforms for robotic applications. TF can be thought of as a live database that keeps track of coordinate frames and their relationships over time. The AICA System leverages this framework internally and facilitates interaction with TF. In particular, user-defined frames are directly included in an AICA application and are available to all components at runtime. These so-called application frames can be created and modified in the 3D scene view, and their poses will be updated in real time.

tip

This guide contains an in-depth review of methods to create, edit, and record application frames in AICA Studio.

Using frames as signals​

In data-flow programming, it might often be necessary to extract the pose of one specific frame from TF in real time and publish that information as a continuous signal.

For instance, a vision component might publish the pose of a detected object to TF, while a motion generator component requires the target pose to be received as a signal. Ideally, the vision component would output the pose as a signal, but this is not always the case. For these situations, the AICA Core provides components that can extract the pose of a desired frame from TF and publish it as a signal, or conversely, receive a pose from a signal and send it to TF.

Frame to Signal​

The Frame to Signal component looks up a desired frame at the specified rate and publishes it on its output as a Cartesian pose. The component has several additional parameters to configure its behavior:

  • Frame: defines which frame should be looked up from TF
  • Reference frame (optional): defines in what reference frame the frame should be expressed in, default is world
  • Remap frame (optional): if provided, the name of the Cartesian pose output will be set to this value
  • Remap reference frame (optional): if provided the reference frame of the Cartesian pose output will be set to this value

Signal to Frame​

The Signal to Frame component receives a Cartesian pose on its input and sends the information to TF at the specified rate. The component has two additional parameters to configure its behavior:

  • Remap frame (optional): if provided, the name of the transform sent to TF will be set to this value
  • Remap reference frame (optional): if provided the reference frame of the transform sent to TF will be set to this value