Pilot Sorting Study

The pilot study will record a ROS bag, but will also hopefully process in real time the video data to generate user level measures for statistical analysis. It will consist of a pool region pieces and three or four sort regions. Puzzle pieces will be placed in the pool region and must be moved to their assigned sort regions.

Each participant has to perform the sorting task several times. All tasks per participant will be a single sitting, from which the breaks must be determined.

The supervisory system will observe where the pieces were placed, keep track of what sort region they were placed in, and the user activities/timing. A marker will be used to signal termination of a study run, with marker removal indicating commencement of the next study run.

Termination perhaps should be done at the comman line (hit 'q' style). The suggested alternative to signal termination is to place a different marker (other color) in same place, or same marker in a different place.

Pilot: Calibration

The core calibration files will be hosted in the Mary (the Puzzlebot) repository, as it is core puzzlebot code. Expected calibration for Mary includes:

• Workscene calibration (black mat, glove, planar workspace).
• Zones calibration (pool and sort regions)
  • First clicked point should be the pool region. Rest sort regions.

The implementation will use the Workscene perceiver, an activity regions monitor add-on, and reporters for the regions.

If run at GT and UCF, then the calibration files should differentiate via GT and UCF prefixes.

Pilot: Execution

To run the sorting pilot involves execution code in the supervise-it repo, within the Pilot_Sorting folder.

./run_sorting_pilot

It will load the calibration files and save the data in the supervise-it data folder (Sorting sub-folder).