Time: 7:38-9:27 Saturday

Weather: low 38.1F (3.4C); high 80.2 (26.8C)

Crew Physical Status: Energetic, focused on the final test

Narrative of Field Mission Results: We began with a thirty-minute meeting to sequence the day's activities, focused on the Scenario 3 test at Lith Canyon. Rick Alena, working at the second repeater and Lith, would tune the repeater relay to achieve a stable computer network signal over the 5 km. Sekou Crawford would drive an ATV to the first repeater a mile south of MDRS to provide a radio relay to Rick. When we received a call back on the IP phone (provided by a kind of server set up by GRC on our network), Maarten Sierhuis would drive the backpacks with minitbooks to Lith; John Dowding and Charles Lee would then walk the backpacks through the scenario, while Maarten and Rick, monitoring from an ATV running DLink (802.11b) searched for blind spots (network dropouts) from the minibooks back to the hab. Ron van Hoof, HabCom, would provide this final information, so we would know when and whether we could run Mobile Agents in the Lith Canyon area.

(Unfortunately, because of the restriction of two repeater hops, the biosensors would not be able to operate using the Cisco network. Rather than reconfigure the iPAQs at this late time, we reluctantly decided to drop the biosensor tracking from the scenario.)

The IP phone call arrived about 12:30, Maarten departed MDRS; Bill, Abby, and Brent then began preparing to drive Everest (the Michigan "pressurized" rover) to the canyon. They arrived just before 13:30 at a quite warm (77F), dry (6% humidity), and breezy site. The walkthrough established that the first third of the canyon was invisible to the Rocky-C ATV, which was distributing the DLink signal. However the middle third, as well as the approach area above, was in good range. The return walk out of the canyon around the bend was visible to the ATV, but too far to detect the minibooks. With this configuration, we could run the planned EVA scenario, but voice commands would queue in the minibooks, to be relayed when the astronauts stepped into range of Rocky-C.

Brent and Abby began suiting up in Everest about 14:45; the minibooks were turned on about 15:00. This established the end of the sim as 17:30, when batteries would be exhausted.

Meanwhile, Nathan Howard and Rob Hirsh drove the ERA (Boudreaux) along the canyon rim to a point near the bend, where it was parked right on the edge. Besides being the most picturesque setting, this location enabled the ERA to serve as a DLink bridge for the minibooks down in canyon. However, this configuration proved to destablize the network because of antenna placement on the ERA and had to be abandoned. (The Mobile Agents software layer integrates Rocky-C with the ERA, but not all functions of the network layer below are compatible.)

The minibooks were turned on inside Everest, but the link to Abby's minibook remained unstable, even after the egress. The Mobile Agents initialization process requires that all core platforms-minibooks, ERA, Rocky-C, and HabCom-be in contact with each other. Standing in the shade outside Everest did not help, possibly because the huge metal reflector of the truck's side, or possibly because Everest, Rocky-C, and the nearby ERA transmitter on a hill were nearly aligned. Also, winds were now picking up, causing the distant repeaters to sway out of alignment.

Consequently, the first steps of the scenario had to be adjusted, so initialization was moved to the first stop, the hill of fossilzed wood (surely known by most MDRS crew members). Here minibook B (Abby's) was restarted and rebooted. (PC Anywhere on Rocky-C was showing a mess of icons on its desktop). Rick believed that the placement of the Platypus water pouch between the backpack's frame and the minibook might be acting as a good insulator, hiding the 802.11b signal. Indeed, the walkthrough test, which had worked at this location, had not included the Platypus pouches. (Furthermore, the walkthrough had neglected to start back at the parking area. This is one of our key lessons, reflected in our caution with "on the bench" testing last week--you might as well set up the full scenario and run it, for that is the system you must get working.)

Finally, about 16:00, an hour after starting the minibooks, and with pressure mounting to begin the EVA, lest the test be scrubbed for the day, Maarten successfully pinged Abby's minibook. Ten minutes later Ron (speaking to Maarten at Rocky-C via IP phone) reported connection back at MDRS and began the Mobile Agents initialization process. Amazingly soon, two and a half hours after arriving at the site, Abby and Brent were following the scenario, naming the location, creating a sample bag, taking photos, and providing voice annotations for these. They then proceeded into the canyon (into the shadow of the Rocky-C DLink relay) and explored the area. Earlier in the morning, Bill had reviewed an aerial photo of Lith Canyon with them, pointing out where they would be working.

During the scenario, they encountered an obstacle that required climbing down, precariously close to a puddle of deep red water, well visited by martian cattle. They elected to take a different route, which brought them to the ledge where Boudreaux was parked (nicely illustrating how the activity tracking system must be extended). Posing and smiling, Brent and Abby asked Boudreaux to take a picture of them. Shortly thereafter Brent reported that his minibook was no longer responding. With battery power exhausted, and the NREN folks waiting for Bill to arrive back at MDRS to participate in a video conference, we elected to not replace the batteries. The long anticipated Lith Canyon scenario, planned nearly a year in advance, was now a wrap.

Successes for the day include: A simulated EVA in full suits and gear was carried out over a distance of 5 km from MDRS, with science data and GPS tracking sent back to the hab, and hence relayed via email to a simulated science team on Earth. Network outtages were frequent, but still enabled the astronaut team to carry out some useful work. Although the usual variety of bugs and timing late in the week precluded the full experience we had hoped to provide to the Northern California Mars Society support team, a great deal was learned about relating the Mobile Agents system to the collaborative work of the astronauts and mission support.

A video conference was held at MDRS, with Bill Clancey, Brent Garry, and Abby Semple seeing and conversing with six representatives of the NASA-Ames NREN team. The conference was held outdoors in the bright sun, lasting about 30 minutes. Ironically, the Brahms (Mobile Agents) group has not worked closely with NREN back at Ames, but here we were in the field with our research projects integrated and meeting each other over the network. Providing such experiences is indeed NREN's mission (see http://www.nren.nasa.gov).

Plans for Tomorrow (Saturday): Pack and clean up.

Report Transmission Schedule: A final report without photos will be transmitted Saturday evening.

Maintenance: We added another 1.5 quarts of oil to the Generac about 19:00. We determined that we would need to switch to the backup gas generator about noon on Saturday.

EVA Narrative/Data/Interpretations: The Lith Canyon EVA scenario, including pressurized rover, ERA, GPS tracking, and science data was first planned last spring, based on FMARS 2001 and MDRS 5 experiences using GPS devices. The idea of tracking a moving ATV between MDRS and Lith now seems naive, as we have discovered that today's networking systems do not allow sustained connectivity over such distances, between hills, with a moving target. Nevertheless, the simulations a year ago provided sufficient experience to focus the Mobile Agent's project.

From Commander's Logbook, April 15, 2002: At this time, I send an email to my colleagues at Ames. I want them to begin thinking about designing software that will make the GPS unit fully invisible to an explorer. I don't want to wait for a satellite fix; I don't want to transcribe readings. I don't want to ever know the coordinates at all, let alone have to manually enter or compare them or number waypoints. I want a program to answer questions while I'm on EVA: "Has anyone taken samples near here before?" I also want the program to tell me things like: "Warning, you are within 10 minutes of the reserve fuel supply required for safe return to the hab." Being here at MDRS this past week has given me very clear ideas about the navigation assistance and other monitoring required during remote exploration on ATVs. Until now, back at Ames and JSC, we weren't sure what to build; we had the methods, but not the requirements. That's why I call what we are doing here "empirical requirements analysis"--finding out what you need to build by doing simulations in the field.

The first designs for the integrated Mobile Agents architecture were sketched at JSC in January 2002. The first field trial was carried out in May 2002 at JSC's "Mars yard"; with the ERA and Rocky-C a few meters apart, it took four days to successfully execute the single voice command, "Boudreaux, take a picture of me."

The team regrouped at Ames in August, and proceeded to a more complex field test for two weeks near Meteor Crater, Arizona. Plagued by the wettest September weather in years, it took more than a week to get the system working. But finally, sample bags were logged with voice annotations, with (periodic) GPS and biosensor tracking with an astronaut riding an ATV within line of sight of the Rocky-C ATV.

In the MDRS16 field tests, we have added a second astronaut, mounted computers on the backpacks, set up repeaters allowing non-line of sight EVA locations, greatly expanded the voice commanding vocabulary, added route and schedule tracking, and provided for emailing data to remote sites. We have lived through the frustrations of technology, the delays, the unpredictabliity of networks, and small, but fatal programming mistakes. (Only one photo was emailed in our Scenario 3 test because the variable REPEAT was not TRUE on the rule that handles emailing. No voice annotatations were emailed because they were handled by the same rule.) Yet when one considers that the idea for building the futuristic EVA system we have demonstrated here was just a glimmer of an idea a year ago, the progress is astonishing. Perhaps the engineers and scienitsts in the 1960s experienced this same rapid and exciting mix of vision, development, and mounting capability.

Inventory: With a hiatus between crews, we've accepted the new challenge of finishing the eggs and bacon.

Miscellaneous: Commenting on the group's progress, Brent Garry said, "Ten days ago, we couldn't get off the front porch. Today I was walking through a canyon 5 km away talking to the hab computer!"