Mars Exploration Rovers Update: A Restored Spirit Studies Adirondack, 'Talks' to Mars Express
Posted by A.J.S. Rayl
08-02-2004 11:00 CST
Opportunity Drives to Outcrop
A fully restored Spirit got back to work late last week on Gusev Crater, using her rock abrasion tool for the first time to brush, then grind into Adirondack. And before her first day back was done, the rover returned more surprises to the science team in the flow of research data she sent home, and reestablished the "international, interplanetary communication network" by exchanging communiqués with the European Space Agency's orbiter, Mars Express.
The plans call for Spirit to hit the road his weekend, heading for Bonneville Crater, the large crater destination selected by the science team.
On the other side of Mars, at Meridiani Planum, Opportunity completed testing and using her suite of tools on a patch of soil near her egress position, and then drove to the outcrop that borders the rim of the crater into which she rolled last month. There, she is beginning her study of Snout, the rock at the very edge of the outcrop.
From Gusev Crater
"I think I can say with as much certainty as we can say anything here that our patient is healed," Mission Manager Jennifer Trosper announced at the Friday morning news briefing held at the Jet Propulsion Laboratory (JPL).
Spirit's computer encountered a glitch in her flash memory that caused her to be unable to carry out her geological fieldwork January 21.
As it turns out, the software specialists assessment that the rover's computer system was having a problem with its flash memory file system was correct and a successful erase-and-reformat operation conducted Wednesday restored the rover to 'like new' condition, Trosper said. The flash memory is used to wake up the rover and to store data over time.
"All indications were that that worked extremely well," Trosper added. "Of course, it was nerve wracking being [93 million miles] 200 million [kilometers] away and reformatting our flash but in the end the spacecraft did exactly what we wanted it to do and it performed perfectly and is in great health right now."
Restoring a rover
The problem was exactly what the primary theory had indicated a week or so ago. Glenn Reeves, flight software architect for the Mars Exploration Rovers, explained exactly what happened:
"In a nutshell, the problem is as simple as this: as we accumulated additional files in the onboard file system, the system consumed additional memory and eventually we ran out. That was problem number 1.
"Problem number 2 is that our reaction to that was fairly severe and in the course of reacting, we corrupted that same file system. The answer to getting back to a state where we are healthy and pristine was to basically get all the information out of that file system that we could, and then to erase the flash memory that holds it and recreate the file system on the vehicle and that's what we did."
The team considered this flaw to be "very severe error," Reeves, added. "The software behaved exactly as we expected it to -- it detected that condition and reset the vehicle. When you go through this action of resetting the vehicle and reinitializing that can clear up the problem completely in many cases. Unfortunately in this case it did not."
After the first reinitialization, the system continued to consume the same amount of memory. "That put us in a situation where the vehicle would go through a rest, then come up and declare an error, and then it would go back down to this reset cycle again," explained Reeves, noting that it took a couple of days to conclude that.
During the design period, the flight software team anticipated a situation like this occurring and so built into the system a mechanism -- the cripple mode -- that allows the ground controllers to bypass going through the file system in flash, and, for example, wake the vehicle in other ways. "On Sol 21, January 24, we were able to get the vehicle into this cripple mode for the first time, which allowed us to stabilize the vehicle, and to get control of the communications again and which gave us the opportunity to go in and 'debug' the system," Reeves recounted.
"The only drawback about debugging a system that's [more than 90 million miles] away is -- imagine your slowest dial-up service provider," Reeves suggested. "Given the distance and the number of times we can communicate, the overall throughput is fairly low.
The team capitalized on every single opportunity it had to 'talk' with the vehicle, and that was, Reeves asserted, "critical" to getting the problem solved as quickly as they did. "We had multiple strategies going on at the same time, and multiple theories in place to make sure we extracted the most information we could."
The actual debugging process involved, Reeves said, "a series of low-level commands and interactions with the system" that enabled them to conclude their theory about being out of memory was correct. "This led us to the point where we were able to basically look at the file system on board and concluded that it appeared to be, for the most part, intact." A couple of days alter, on the January 27, they attempted to downlink a lot of the science data products that had been in the flash memory. "And we tried to establish our strategy for recovery," he said.
It was then, Reeves continued, that they realized the system was probably corrupted. That meant we had two problems effectively. . . [and] that led us down the course [of having] to completely reinitialize the file system. The manner in which we chose to do that was to erase all the flash to make sure all of that worked correctly, then go through the process of creating the new file system," he explained. "It is not dissimilar from the way you would attempt to do this same mechanism on your home PC, when your [hard disc] becomes so corrupted it can't be used."
The investigation is ongoing, Reeves noted. "In some of our lander's tests possibly an inkling that this problem was out there, but we ran the system right up to what we considered the limit on the number of files we would ever produce. And I think we ended up with a situation where the system was asked to perform more than we had tested, but we had the expectation that it would have behaved correctly. So this is definitely a bug in our court that we have to fix. The problem itself is actually the configuration of the vehicle -- that we allowed it to consume the memory space as it was needed - the bug, in my opinion, is we should have restricted that."
Reeves said that he and his team are " very confident" that this will not happen again. "Now that we understand what the problem is, and we have a procedure in place, we believe we can work around this problem, indefinitely if we have to." Suggestions for changing the flight software so the glitch could be identified more quickly, he added, may be forthcoming. "But I think both Spirit and Opportunity -- because we can see the issue now -- we shouldn't have any problems in the future."
Back in the Saddle
On Thursday, Spirit was woken up at her normal time 9 a.m., to screeching rock strains of Aerosmith's "Back in the Saddle Again," Trosper said. Once mission controllers confirmed her flash memory was stable and available for data storage, the robot field geologist was cleared to conduct the day's science activities.
Before the day was through, Spirit had logged another milestone on the MER mission -- she rotated her turret and deployed her rock abrasion tool -- called the RAT -- and used its brush to dust-off the rock Adirondack. Then, she snapped images of the result with her PanCam, and zoomed in with her microscopic imager (MI) to take pictures up close. The Mössbauer spectrometer and alpha particle X-ray spectrometer (APXS) took measurements overnight on the same, brushed area.
It was the first time the RAT on either rover had been used and it was the first time a rock on another planet had been "brushed" to remove dust and allow inspection of the 'clean' surface. The brush swirled for five minutes, said Stephen Gorevan of Honeybee Robotics, lead scientist for the rock abrasion tools on both rovers. But the surprise was still to come.
The RAT's primary function is to grind off the weathered surface of rocks with diamond teeth, but the brush was added to clear the grinder's cuttings -- can also be used to sweep dust off the intact surface -- and sweep they did. The tiny stainless steel bristles, set about 3 millimeters below the grinding teeth, cleaned a distinct circular patch on Adirondack.
"We decided to take a brush to Adirondack to remove loose material, but to preserve a cemented coating," Gorevan explained. "Many people in the science room didn't even think we'd see a difference, and I didn't expect much of a difference, but this is a big surprise," Gorevan said as he presented the picture taken by the microscopic imager.
The brushed area on Adirondack appears as a dark circle, much darker than the rest of the rock's surface. In trying to capture the emotion of the team upon seeing the image, Gorevan said: "All I could think of was Muhammad Ali -- Ladies and gentlemen, I present you the greatest interplanetary brushing of all time!"
Interestingly, one reason scientists first selected Adirondack for close inspection was because it appeared from PanCam images taken remotely to be relatively dust free compared to some other rocks nearby. "To our surprise, there was quite a bit of dust on the surface," said Ken Herkenhoff of the U.S. Geological Survey's Astrogeology Team, lead scientist for the rovers' microscopic imagers.
The examination has just started, but Adirondack, Herkenhoff said, "looks like is a volcanic rock, basalt probably." From the MI pictures alone they can see, he added, "what appear to be mineral crystals on the rock surface." The data from the other instruments should be streaming in over the weekend.
Spirit packed another major accomplishment under her exploration field-belt on her first day back at work by successfully relaying and receiving communication from Mars Express as it orbited some 300 kilometers directly overhead. "We actually did a demonstration with Mars Express for both forward link commanding and return telemetry, and it worked very well," said Trosper. "We saved a file on board the spacecraft that was sent to us from Mars Express, and they have telemetry that they have sent to us and we are decoding that. So we have our international interplanetary communications network established even more so at Mars."
On the road again
Yesterday, Spirit successfully deployed her RAT and abraded Adirondack. It was the first time, again, in history that a rock on another planet was abraded. Scientists and engineers were ecstatic when the afternoon communications relay from Mars Odyssey revealed a round and clean-surfaced depression.
On the agenda for today, Spirit is being directed to inspect the newly exposed ancient rock material with her Mössbauer spectrometer, MI, and APXS before making a 20-foot (6-meter) drive around the south side of the lander. Current plans may keep this rover in drive mode for the next few sols as she heads northeast towards a crater the team dubbed Bonneville, about 820 feet away.
"We're on the west, southwest side of the lander and we're going to drive around the lander and head northeast for Bonneville Crater," Trosper said. "We do believe that we will be able to get there. We're having some strategy discussions now on how we would use the capabilities of the vehicle, in terms of traversing and auto navigation, in order to get us there."
Right now, they will design the rover's traverses based on visual data from the images and will give her specific commands and directions. "Then at the end of that traverse, we will turn on our auto navigation software and allow the rover to determine how to get to the next way point, for a meter or so, just to check out the auto navigation software," Trosper offered. "We'll start in baby steps and in each sol we're driving, we'll probably expand the numbers we are using for auto navigation, as well as the distance that we might actually designate as a traverse."
The robot geologist will make study stops along the way, Trosper said, and once at Bonneville will study rocks thrown outward by the crater-forming impact. "There's going to be a lot of driving on Spirit and we've worked with the science team and talked about four or five places we might stop and do specific studies with the IDD and that's our plan," she added.
From Meridiani Planum
Meanwhile, on the other side of Mars, Opportunity continued "to turn the crank," Mission Manager Matt Wallace reported. The second Mars rover carried out a number of tests and studies with three of the instruments on her robotic MI, the Mössbauer and APXS spectrometers, and then drove toward the outcrop that rims the crater, and to the rock at its very edge, Snout.
Opportunity began her jaunt Thursday, driving about 11 feet (3 -1/2 meters). "The drive was composed of several arc turns to the left, followed by an arc turn to the right, a turn in place and then a small drive forward -- all of which were designed to give us additional confidence and characterization of the mobility system and its performance in this crater in Meridiani," said Wallace.
The rover continued the trip, driving to within about 20 inches (half a meter) of the rock Snout on the northeastern end of the outcrop Friday, and completing the final few inches Saturday. Why Opportunity came up short of Snout "is probably due to slippage in the soil," Wallace said.
"We're not entirely sure why we fell short, but we're pretty sure that what we're seeing is soil slippage," said Wallace. "Everything's pointing to the idea that we're getting anywhere between 10 to 20 percent of slip during these traverses. Once we really better understand what this soil is and how the vehicle reacts to it we'll be able to accommodate that pretty efficiently."
Prior to that final little drive yesterday, Opportunity was scheduled to conduct what is called a ' touch 'n go.' "That's where we'll deploy the IDD [instrument deployment device] and drop the arm toward soil and take several more MI pictures to continue to catalogue the soil inside the crater," explained Wallace. "Then, we'll restow the arm and drive 30-40 centimeters and prepare for a full suite of instrument arm activities on Snout, MI, APXS and Mössbauer," he said.
"From there, it looks like the scientists are asking us to start an arc along the bottom of the outcrop area, stopping at several choice viewing locations and taking some addition pictures, and potentially dropping the arm down and getting more MI pictures as we go," Wallace added.
Opportunity has been performing her duties without any trouble, Wallace reported. "She continues to operate nominally. She's healthy and happy. And continued to do the job she was sent to do."
The main task for both rovers is to explore the areas around their landing sites for evidence in rocks and soils about whether those areas ever had environments that were watery and possibly suitable for sustaining life.Each Martian day, or "sol," lasts about 39 minutes longer than an Earth day.