Introduction:

The small Mars crew will undoubtedly have to collaborate with groups of scientists back on Earth. How this collaboration will happen is largely undefined. This year we will test our first design and implementation of a computer supported Mars- and Earth-based science work system. A computer supported work system is a system that includes both the human work processes and the computer tools and dataflow management systems. It is here that human-centered design meets work process design. In the empirical design approach we are using in the Mobile Agents project we let ourselves be guided by the capabilities of the people part of the mission. In other words, the people are at the center of the total system, and the people are supported in their work by our computer tools. We strive to automate that what makes most sense, developing automation that supports the human work practices, allowing people to do what they are good at.

We start simple, asking basic questions like how the Mars crew can communicate their daily EVA plans and captured science data during and after an EVA back to their colleagues on Earth. By starting there we immediately ask ourselves what the role of an Earth-based science team is. Can the remote science team (RST) participate in the planning of daily EVAs on Mars? Will the RST be able to get the science data in time to make useful suggestions to the crew? Will the crew be able to absorb the RST suggestions in a timely manner to develop a daily EVA plan? How will the RST EVA plan compare with the crew plan?

This year we have defined a fairly simple science work process and use three pre-existing software tools that have been used in other domains. We have integrated two of the three tools with our Mobile Agents Architecture (MAA), enabling the automatic flow of data between Mars and Earth. With this integration we show how easy it is to add new software tools to the Mobile Agents Architecture, allowing previously developed software tools and systems to interact with software systems already in the architecture without having to develop dedicated software interfaces between them. This important feature of the MAA allows us to integrate those systems that are developed separately to support specific functions in the work process.

We briefly describe the make-up of the RST and the crew/RST workflow process. We then describe how the three tools support the crew and the RST in doing their work.

The Remote Science Team:

The RST of Rotation 29 of the MDRS consists of a number of distributed scientists from all over the world. Figure 1 shows a graphical depiction of the RST organization. The scientists are divided into two teams, the Mars Society RST and the State University of New York at Buffalo RST. The Mars Society RST consists of a group of four volunteer scientists who are all members of the Mars Society, and have supported all the 2004 MDRS rotations. The members on the Mars Society RST are located in California, Arizona, Canada and Australia. The SUNY Buffalo RST consists of a group of six geology Ph.D. and graduate students from the Geology department at SUNY Buffalo. This Buffalo RST team is co-located in the Volcanoes Studies Group's laboratory. Each RST team has a team lead who is responsible for the working of the team. The complete RST is lead by the RST Leader who oversees the work of all the RST subteams and manages the communication between the crew and the RST.

Figure 1. The Remote Science Team - Click for Detail

The RST teams need to collaborate together to a) analyze the EVA science data send to them from Mars and b) provide feedback on the crew's EVA plan for the upcoming day. This year we defined a simple work process in which the crew plans and executes their EVAs in a daily EVA-cycle. The Earth-based collaboration between the distributed RST members needs to equally be facilitated by collaboration tools. A human RST facilitator facilitates the Science Operation Working Group (SOWG) web-based teleconferences between the RST teams. The workflow process is described next.

The Crew/RST Workflow Process:

The workflow process is divided into two parallel processes that are dependent on each other. Figure 2 shows a high-level workflow process model.

Figure 2. The Mobile Agents Project Crew/RST Workflow - Click for Detail

The process above the dotted line represents the steps for the crew on Mars, while the process bellow the line is representing that of the RST. The process starts on the left side of the diagram showing the crew performing their EVA. This represents a random EVA that has been planned the day before and was briefed by the crew the morning of the EVA (which is depicted on the right sight of the diagram). During the EVA science data captured by the EVA astronauts and/or the ERA robot is both communicated to the hab and back to Earth by the Brahms software agents. The data is stored in both ScienceOrganizer in the hab and on Earth, as well as in the Compendium database in the habitat. The RST members will receive e-mail from the Habitat e-mail agent with a URL reference to the stored science data in ScienceOrganizer. This way the individual RST members can follow the EVA as science data is communicated to and stored back on Earth, allowing the RST members to start their individual analysis as soon as they can.

After the EVA, the crew will analyze the collected science data (rock and soil samples) back in the habitat. After this analysis, the crew will discuss the results of the day's EVA and plan the EVA for the next day. This meeting is facilitated with the Compendium meeting capture software, and is video taped. The result of this is communicated to Earth. On Earth, the crew's meeting video and Compendium database is used to create a MeetingReplay Tool. This allows the RST to view the EVA planning discussion of the crew and at the same time view the crew's plan in the Compendium tool. After the individual RST members have viewed the video, the will come together in a remotely facilitated meeting on the web. The meeting is facilitated using the Compendium tool displayed on every RST member's computer screen over the web. This is called the Science Operations Work Group meeting. The name is taken from a similar type meeting held during each Sol on the Mars Exploration Rover (MER) mission at JPL (http://origin.mars5.jpl.nasa.gov/home/). During the two hours SOWG meeting on the MER mission, the science team decides the plan for the next day. Similarly, the RST team meets together in a teleconference to review the plan made earlier by the crew on Mars. The meeting is captured in the Compendium meeting capture tool. The result of the meeting is e-mailed back to the crew, before the crew wakes up the next morning. The next morning at 8:30am Mars time the crew debriefs the RST plan in Compendium, and decides on the final plan for the day's EVA.

Crew/RST Collaboration Support Tools:

For a description of the tools being used in our experiment go to the RST Communication System Report of 4/27/2004.