Introductory Note:

The RST Communication System Report (RCSR) has been introduced to provide updates on experiments with new internet collaboration tools to support the Remote Science Team, as they work with each other and the Hab Crew across multiple locations and timezones. This will complement the EVA Communication System Report (ECSR) and the RST Report.

This MDRS mission fabric introduces a new strand of research on how best to support the Remote Science Team (RST), who in actual missions are likely to be spread across the Earth in multiple timezones (it will be impractical to physically co-locate all relevant experts who might be needed in a multi-year mission to Mars). Looking ahead 20 years, this research will prototype and evaluate tools to support scientific teamworking under such circumstances, using information networks of the future.

This is the rationale for a new collaboration between NASA's Human-Centered Computing research group and the UK's CoAKTinG Project (part of the Advanced Knowledge Technologies consortium), who are developing new kinds of collaboration technology to support 'e-Science'. This MDRS mission is providing a testbed for some of the CoAKTinG tools, to see how well they can connect the RST with each other, and with a crew located on another planet.

Status:

Since last November, the RST has been holding weekly planning meetings via teleconference, putting the new collaboration support tools (summarised below) through their paces. We are beginning to understand the role that different collaborative technologies can play in the overall work ecology of the RST and Hab Crew.

The RST has been using the ScienceOrganizer system to archive and to link meaningfully their project materials, and experimenting with Compendium to design their own maps to support a geological methodology. A trial version of the Meeting Replay Tool from the Operational Readiness Test also generated much interest, and was fine-tuned as a result. The BuddySpace messaging and awareness tool will shortly be released for trial. Below we briefly summarise the different RST Communication System tools.

ScienceOrganizer:
long-term, formal memory for mission data and knowledge

ScienceOrganizer was developed at NASA Ames, and provides a Web-based semantic database that can be populated and searched by scientists or engineers. It has been previously used to support the Columbia Shuttle accident investigation, to study patterns of malaria in Africa, and to design and execute experiments on microbes, among other applications. ScienceOrganizer provides a rich interface for semantically linking resources, and is most suited for long-term knowledge management, especially for distrbuted teams collaborating asynchronously.


A ScienceOrganizer database entry showing an image taken by an "astronaut" on a simulated EVA.. On the left are hyperlinks to support browsing of resources related to the photo, such as the location the image was taken, the astronaut, and the EVA itself.

As the ScienceOrganizer interface consists of web pages, the system is available to virtually anyone, anywhere. However, it was not designed to support synchronous, real-time collaborations, or activities such as meeting capture, free-form thinking, and brainstorming. One of the questions we are investigating concerns how this system could complement another, introduced below.

Compendium:
Mapping meetings as group memory

Meetings are as pervasive an activity in scientific life as elsewhere, but similarly, can often also be ineffective. Moreover, when conducted in an international context via a virtual environment such as a video- or tele-conference, many of the normal non-verbal cues that add fluidity to interaction are missing as well. The Compendium tool being developed at the Open University provides a way to focus discussions on specific topics, integrating other documents into the idea network that grows as contributions to a discussion are mapped on screen. Compendium differs from other concept mapping tools in the specific visual language it provides to capture and organise issues in real time as they emerge in discussions, and in the hypertext and semantic tagging mechanisms it provides to manage large databases.

Compendium will play several roles in this mission.

1. Firstly, it mediates between the Hab Crew and software agents. The Crew will use it to plan the route of the next day's Extra-Vehicular Activity (EVA), constructing a visual map of the locations they want to visit, and the activities to be conducted at each. These Compendium maps will then be 'read' the next day by the software agents which coordinate the human and machine elements of the MDRS infrastructure during an EVA.
   
   
   An EVA plan constructed by the Crew in Compendium, subsequently read by the Brahms software agents which coordinate human-machine collaboration on the EVA.
   
   
2. Next, it can be used to view all the data gathered on the EVA when the Crew returns to the Hab. All the data generated during an EVA (videos, photos, voice annotations...) stream back over the wireless network and are stored in the Hab in ScienceOrganizer and Compendium. The latter displays data in concept maps, assisting analysts in seeing and navigating certain kinds of structure, and linking in their own ideas and documents. All Compendium data is also archived in ScienceOrganizer.
   
   
   Following the EVA, a map is created and populated in Compendium rendering the data gathered. Shown is an example Science Data map, highlighting the data gathered by Astronaut One.
   
   
3. Finally, it is used by the Crew and RST to discuss the implications of the data. Compendium displays information in visual concept maps in order to support sensemaking activities such as planning, discussions, and the gradual structuring of ideas. As the discussion unfolds, the contributions are simultaneously mapped on the screen (projected in the Hab in a Crew meeting, or shared over the internet in an RST telecon). The RST can therefore see what options the Crew may have discussed but rejected, and why, in order to provide scientific feedback to the Crew.
   
   
   Compendium map summarising the RST's feedback to the Crew. Each node is hyperlinked to the map (i.e. the context) in which it was created.
   
   The insight that the RST gains into the Crew's decision-making is further enriched by an indexed, digital video record of the Crew's meeting which integrates Compendium with other resources, as described next.

Meeting Replay:
Enabling the RST to review a Crew meeting

With a 20 minute communication delay between Mars and Earth, the usual ways of working together at a distance, such as conversation and the sharing of computer screens, are impractical. As part of the CoAKTinG project, the University of Southampton are developing a meeting replay tool, which combines meeting materials within an interface structured to enable quick and easy indexing navigation of the meeting record.

During the mission we are recording the Crew's daily EVA planning meetings and delivering a replay of the meeting to the RST within a few hours. By experimenting with these techniques we hope to see if the RST can gain a better understanding not only of what a Crew is deciding, but why, and how, in order to provide the best kind of feedback.

In the screenshot below, you can see the Web-based Meeting Replay Tool. The upper region shows the video of the meeting and the Compendium map as the discussion progresses. The lower region contains summary information about the meeting - who was there, who was speaking, the agenda, and an overview of the current topic (derived from the Compendium map). Some of this information is presented as a timeline, which gives a visual index an RST member can use to navigate the video, jumping to relevant or interesting parts of the discussion by clicking on the timeline or moving the slider.


When reviewing the meeting replay, Compendium has been extended so that it can be used as a 'visual contents page' into the video. For instance, if the RST wants to see discussion prior to the recording of a particular Decision, one can now click on this node in Compendium and the replay jumps to the point in the meeting where that node was recorded, and starts to play the video.

When the RST meets collectively, each participant runs their own meeting replay. Should the need arise to review a specific point in the meeting as a group, one member can take control of all the meeting replays using the GroupSync feature. As that member navigates around the meeting everybody else's client follows their lead, remaining in synchronisation as long as is required.

BuddySpace:
Instant messaging meets presence visualization

In a team distributed in time and space, staying aware of colleagues' activities and availability makes up for the the reduced sense of 'presence' that we have compared to when co-located. Our challenge is how best to characterise presence, how to make it easy to manage and easy to visualise, and how to remain consistent with the user's own expectations, work habits, and existing patterns of Instant Messaging (IM) and other communication tool usage.

A prototype presence management tool called BuddySpace is being developed at the Open University, and overlays presence information onto visualisations, both geographical (e.g. a map of a building, region or planet), and conceptual (e.g. a workflow chart, project plan, or experiment). In the RST BuddySpace map below, the availability of members is shown mapped to their geographical location.


Prototype RST BuddySpace map, in which the online availability of colleagues is superimposed on locations in the US and UK. Clicking on one of the presence dots opens a conventional instant messaging session.

(On a technical note: BuddySpace is based on Jabber/XMPP, and is already interoperable with MSN Messenger, Yahoo, ICQ, and AIM -- we plan to deploy BuddySpace once the mission is underway and the above tools are all working together.)

Our experiences to date suggest that together, these are a potentially powerful suite of tools for supporting RST-Crew collaboration, but of course only hard experience can tell us as we deploy over the next two weeks.

RST Communication System Team

• NASA Ames
  (ScienceOrganizer: Dan Berrios, Rich Keller)
• The Open University
  (Compendium: Simon Buckingham Shum, Michelle Bachler & Al Selvin)
  (BuddySpace: Marc Eisenstadt & Jiri Komzak)
• University of Southampton
  (Meeting Replay Tool: Danius Michaelides, Kevin Page, Dave De Roure & Nigel Shadbolt)

We gratefully acknowledge the UK's EPSRC who fund the Advanced Knowledge Technologies consortium, of which the CoAKTinG Project is part: www.aktors.org