Crew 38 returns to MDRS to conduct field experiments using the wireless EVA data management system ("Mobile Agents") that we have developed over the past two years. We previously showed in Crews 16 and 29 how to use MDRS as a research station-a protected workplace for scientists and engineers to configure, deploy, and test sophisticated technologies and protocols in an extreme environment. As in previous years, only the EVAs are conducted in simulation mode (with two astronaut-geologists in suits wearing computer backpacks, and a support crew of 10 or so people off camera).

The project name, Mobile Agents, refers to software running on moving computers-those carried by the astronauts and the rovers. To answer these questions and carry out commands, the astronaut computers, the ERA computers, and the HabCom computer in MDRS are networked and communicating information as required. Over twenty researchers are collaborating from two NASA centers and several universities. (See Maarten Sierhuis' Mobile Agents Architecture essay from Crew29 for details.)

Last year, we connected a database ("ScienceOrganizer") to the network in the hab and made it accessible to a remote science team (RST). We carried out very successful, long-duration tests in a canyon 5 km north of MDRS. A combination of tools allowed the crew and RST to collaborate in reviewing robotic reconnaissance data and to plan EVAs. Astronauts recorded voice notes, took photos, labeled samples, and directed a robot to take panoramas-with all of this data transmitted automatically, time and place-stamped, and stored in ScienceOrganizer, and emails sent to the RST providing reports of the EVA status and pointers to the data.

This year we are adding a second Extravehicular activity (EVA) Robotic Assistant (ERA), with a capability to automatically deploy a wireless network repeater station (robotically deployed relay, RDR). Plus the robots will have a rudimentary capability to explain to the astronauts what they are doing (e.g., "Who are you following?" "Where are you?") and the state of the network connection ("Do you still have comms?"). In addition, the astronauts will be able to ask their "personal agents" (software programs running on the computers they carry) basic navigation questions (e.g., "Where is the next activity?" "Where is astronaut one?").

The astronauts have about 50 categories of commands for affecting the GPS, biosensors, EVA plan, and science data (images, voice annotations, and samples) -all during the EVA. They can even print labels for sample bags from a printer on the ERA's equipment cart. In conventional terms, we're providing a voice-operated "work flow" system. The created data is routed, copied, and stored appropriately; plus the Mobile Agent system is keeping track of time and the astronauts' location and health, and providing warnings verbally to the astronauts, as well as on the loudspeaker in the hab, and via email to the RST (e.g., "Astronaut one has 5 minutes to complete the work site two activity.").

This year we intend to use the Mobile Agents system in a less-scripted "test" mode, and shift to experimental usage that is more improvised. We will begin in the first week by a pedestrian EVA to the "Pooh's Corner" area, to the east across from MDRS, to do a shake-down test of the new robot hardware, commanding, and automated robot behaviors. First, an ERA will do "autonomous" reconnaissance, measuring the adequacy of the wireless network (by moving in a spiral or rows back and forth) and creating a rudimentary map that will be used for EVA planning. The ERA will automatically deploy a wireless relay (RDR) to enable it to extend the reach of its reconnaissance. After a day of planning with the RST and training the astronauts, the astronauts will walk to the planned locations and perform a survey documentation of sampling, photography, and description. The second week will be further to the east, requiring a combination of a backbone relay from MDRS and up to two relays. Depending on requirements discovered in the field, the astronauts can direct the robots to deploy an RDR or to stay in a certain spot (such that the ERA serves as a relay). If all goes well (weather and equipment permitting), we hope to perform two such explorations, replanning each based on the previous work conducted in the area.

The research we are conducting concerns how to build an efficient and effective human-robotic system. Although the experimental scenarios may appear to focus on network relay functionality (providing coverage back to the habitat for the robots and astronauts), this is not the focus of the research, and not how we are suggesting networks will be best deployed for planetary surface exploration. Rather the network relay coverage is used only as an example service that robots could provide that exercises fundamental aspects of human-robotic interaction (HRI). The important point here at MDRS is that the need for the service is real, and successful interactions with the robots are essential for the astronaut's exploration work.

Crew 38's Commander (CDR) "check in" reports will provide brief updates on the status of system deployment and testing. Fuller reports of EVAs will be written by Brent Garry and Abigail Semple, our geology-astronauts. Maarten Sierhuis, Mobile Agents Project Lead, will report on the interaction of the crew with the RST and other aspects of the Mobile Agents field test in an "EVA Communications Systems Report." Engineering status will be reported by Frank Schubert; Greenhab by Liam Pedersen in the first week. Abby is also the Health & Safety Officer.

Bill Clancey
April 3, 2005