During July of FY16 we were notified that Playbook is slated to be used onboard to evaluate crew autonomy as part of the CAST (Crew Autonomous Scheduling Test) technology demonstration task onboard the International Space Station. The CAST technology demonstration, which is expected to start at the end of December 2016 and extend through early 2017, involves 5 different objectives to evaluate and test crew autonomy. The CAST demonstration will involve the crew planning and executing off of their plan at different levels of granularity using the Playbook tool.
As part of this effort we trained Astronaut Peggy Whitson on the Playbook tool and have been working closely with our FOD (Flight Operations Directorate) counterparts to prepare for this new technology demonstration. The goal of CAST is to inform policy and process decisions on crew autonomy for future deep space missions.
In Year 2 we completed development of the Playbook Data Analysis Tool culminating in our first version that incorporates the playback and analysis features. With this tool we are able to playback the unobtrusively collected use data (clicks, gestures, etc.) from the analog missions in a video tape like fashion.
The interface includes a playback interface for the session, allowing you to click to seek to the exact time you’re interested in, play, and pause the playback. Below the playback controls there are individual timelines that show the frequency of usage of specific views (such as the Timeline, Mission Log, Procedures, etc.) or formulaic queries throughout the self-scheduling session. For example, one query that may be called “view=Timeline”, will show me usage of the Timeline throughout the unobtrusively collected session.
New queries can be added and existing queries can be edited on the fly. For example, if you were interested in sessions when the users were using the tablet in a portrait orientation but then wanted to refine that to only show the transitions from a portrait orientation to a horizontal orientation, you can update the query inline, which will render a new timeline for that result. This allows users to tweak and compare their queries as they are performing their analysis. When a user finishes adding the queries they would like to visualize, these information metrics can be exported as part of the CSV (comma-separated value) export, as well as the original raw unobtrusively collected data.
Playbook continued to be deployed and used as the primary operations tool for the HERA analog in year two of this proposal effort. The HERA crew and operations staff used Playbook to communicate and simulate communication time delay, status activity plan execution, view procedures for experiments, and to status and review the mission plan. It was deployed and used as the primary mission planning and text communications tool for all four HERA Campaign 3 missions within year 2 of the proposal effort.
In year 2 we were able to have the crew perform limited self-scheduling of actual mission activities and were able to fully collect unobtrusive data and user surveys, in contrast to year 1. In addition to that we were able to conduct several self-scheduling exercises to exercise and evaluate the usability of the features in the analog. These self-scheduling exercises are done using a separate fictional timeline and allow us to look at more complex self-scheduling problems as well as thoroughly evaluate the usability of the self-scheduling features. We are currently performing an initial analysis of this data as Campaign 3 wraps up.
Playbook was used on NEEMO 21 in support of all four of our proposal aims. We were able to gather input and evaluate all features created for a future ISS deployment, supported NEEMO 21 operationally using Playbook, evaluated crew self-scheduling in an operationally realistic setting, and collected click and gesture data using our unobtrusive data collection functionality which we built into the Playbook tool as part of the proposal research effort. We were also able to give the crew surveys on Playbook use.
This year on the NEEMO 21 mission, we had two main iterations of our crew autonomy research. Following up on our work last year, we were trying to address the issue of the cost benefit analysis of crew self-scheduling. Specifically, we were trying to figure out if we could transfer some of the more tedious tasks of self-scheduling from crew to ground to lower the overall amount of effort involved with self-scheduling. This was done in one of two ways: strategic and tactical planning. The idea was to further abstract the more tedious aspects of planning that are not necessary to be done by the crew, such as fine tuned placement of activities and manipulation of small activities, and instead focus on allowing the crew to express their overall intent. So with the strategic planning concept, which was derived from the idea of “Plan Suggestions” mentioned in the previous annual report, we allowed the crew to plan their EVA (extravehicular activities) using a new feature we introduced this year into Playbook called “Add Activities.” This feature lets you freely add, edit, or remove crew-created activities into your mission plan. For the tactical planning concept, we updated the exercise that we conducted last year with a couple of new features, the ability to have groups in the task list and the ability to move groups to and from the Scratchpad. In both strategies the crew planned out two days of their EVA activities. These concepts were executed separately with the strategic planning concept taking part in the first half of the mission and the tactical planning concept taking part in the second half of the mission. It should be noted that there was a partial crew change out during the NEEMO 21 mission so only 2 crewmembers experienced both planning strategies.
The latest version of Playbook includes the ability to link to ISHORT procedures if the activity has an associated ISHORT procedure. This feature was built and deployed out for the NEEMO 21 analog; however, since ISHORT was not on NEEMO 21, it is yet to be tested in an analog environment. We tested the linking behavior internally in the lab environment during development.
|