A Feasibility Analysis of Primitive Level Task Allocation for a Cooperative Human and Robotic Team

Paper number

IAC-08.B3.6.6

Author

Ms. Sharon Singer, Space Systems Laboratory, University of Maryland, United States

Coauthor

Dr. David Akin, Space Systems Laboratory, United States

Year

2008

Abstract

Future space missions will benefit from leveraging the unique capabilities of a cooperative human and robot team for performing mission activities.    While robots specialize in tasks requiring the repetitive use of a single axis tool, humans perform fine dexterous work and manipulation much better.  A fundamental aspect of planning an efficient schedule for a combined human and robot team is allocating the set of tasks to optimize the time that humans are involved.  To facilitate task allocation, mission activities can be broken down into three general categories: tasks, subtasks, and task primitives.  Primitive level task analysis records the tools and movements that are required for a given task.  From this level it is readily apparent whether a team member has the capacity to perform a required task.  Allocating tasks at the primitive level for future human and robotic missions would yield the minimum active involvement time of the crew in any activities.

This research effort sought to assess the feasibility of primitive level task and its subsequent impact on team performance.  Time durations for task primitives were recorded from both the flight log and flight footage from the Hubble Space Telescope Servicing Mission 3A (HST SM-3A).  The intention was to apply scheduling algorithms to generate alternative timelines for a human and robotic team from the primitive data.  Application of this method was impeded by the incompleteness of the two data sets in representing the tasks performed by the HST SM-3A crew.  Neither data set was found to be more consistent.  The quantifiable benefit of using both data sources was also assessed.  To improve the representation of the data set, task primitives were combined into batches based on properties inherent to the tasks (tools or interfaces used, degrees of freedom necessary to perform the actions, etc.) while maintaining precedence constraints.    

Depending on the types of tasks required in future space exploration missions, allocating task primitives within a human and robotic team could enable the crew to perform a larger number of tasks within the same time period than a purely human crew could.  However, workspace obstruction issues and crewmember wait-time resulting from poor batching of the primitives decreases the utility of the cooperative timeline.  The performance of cooperative activities by humans and robots in future space missions will be greatly enhanced by batching similar primitives during the task allocation and scheduling of the mission timeline.

Abstract document

IAC-08.B3.6.6.pdf

Manuscript document

IAC-08.B3.6.6.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.