Pose Training and Run-Time Alignment

To accurately align robot motion with detected parts, In-Sight Spreadsheet uses a pose training and run-time alignment process based on vision tool results.

During setup, train the robot arm pose using a representative training part before runtime. Use a fiducial marker or identifiable feature on the training part that is consistent across all parts, such as a fixture A fixture is a coordinate location on the image that keeps the tool in the place determined by the fixture. mark, edge, or pattern, as a reference for alignment.

Using standard vision tools and the Robotic String Functions, the vision system registers the pose of the robot arm relative to the fiducial marker on the training part. At runtime, the vision system applies the same relationship to each loaded part by locating the fiducial marker and calculating the corresponding robot target pose.

The typical pose training and alignment workflow proceeds as follows:

1. Configure a vision tool to locate a fiducial marker or distinguishing feature on the training part. For example, if all parts contain a consistent feature in the same relative location, you can use a Pattern Match tool as the alignment reference.

2. Perform hand-eye calibration to establish the relationship between the vision system and the robot coordinate system.

3. Train the robot arm pose relative to the fiducial marker.

4. Repeat the previous steps for every robot arm pose that you need in your application.

Once training and alignment is complete, the system executes the following workflow during run time:

1. The vision system locates the fiducial marker on each part.

2. The vision system computes the required robot pose based on the trained alignment.

3. The vision system sends the resulting pose to the robot controller.

4. The robot controller moves the robot arm to the calculated pose.