2009-04-25

Spirit Sol 111

Once again, things have gotten a little too routine, so we've been asked for something unusual. They want to MTES a couple of nearby rocks, but we're not close enough to do it from here. So we're going to drive about 12m to a position near both rocks, MTES them from there, and drive on.

And this wouldn't be a problem except that the MTES's field of view is so narrow. The rover might (read: will) slip while traveling those 12m, which causes it to lose track of its position and thus to become unable to accurately aim its instruments. But we don't know how badly it will slip, so we can't correct for this in advance. It's like hitting a pinata blindfolded -- when the pinata is the size of your thumb. And you only get one swing.

Robin Fergason, the MTES PUL for the day, is just as worried about this observation as John and I are, so she and I spend some time looking at the rover's geometry, ensuring that the rover's own body won't block the target. It's a good thing we did so: the left solar panel will almost obscure the rock; if our simulation is off more than a little, we'll waste our shot. We could turn the rover a little to the right so that the rock will be visible through the "notch" between the left and rear solar panels -- but then, the HGA gets in the way. To heck with that: we end up adding commands to turn the rover to face the rock.

This isn't the first time we've had a concern about hitting a target with the MTES. The combination of slip and the MTES's narrow FOV has made this problem particularly acute for that instrument. Indeed, this problem goes all the way back to our drive to White Boat -- my first solo drive. But Robin reassures me that we've never missed one yet. There have been a few partials, but we've always come close enough for them to get the data they were after.

Thanks to my perverse nature, this makes me more worried about this drive, not less. Now that I know we have a perfect record, I really don't want to spoil it.

The rover does have a feature, visual odometry, which will let it notice and correct for slippage. But that's much slower -- it will take half an hour or so out of the day's drive -- so when Robin and I pitch the idea to Squyres, he nixes it. The observation's not important enough to justify that cost, he says. "Just take your best shot."

So that's what we end up doing. John works out a path that will avoid large rocks (which contribute significantly to slip) as best we can, and then we just cross our fingers. We've done all we can.

I really, really don't want to blow our perfect record, though.

At the end of the activity plan approval meeting, Art habitually asks the same question of the rover driver, the TAP, the SOWG chair, and others: "Good plan? Go forward?" When he asks Squyres that question today, Squyres shrugs: "I wouldn't put it in my top 10 sols of the mission or anything, but it's OK." "Mediocre plan, go forward?" Art asks. Squyres laughs and agrees.

We're finished and leaving almost before I know it. The day's planning started at 6:30AM (I came in at 9AM; it's so much nicer to be the later shift), and we're officially done at 2PM. Everyone else loves the compressed schedule, but I'm not a big fan. We spent every sol of the nominal mission planning the rover's day as efficiently as we could, and we took the time (as much time as we possibly could) to do it properly. On the compressed schedule, we're not doing that any more -- which means we're not maximizing the value of the rovers. This bugs me, but there's no solution short of returning to Mars time, and there's no hope of that. (To be fair, switching to Earth time was a money-driven decision as much as anything else: there's less money to run the extended mission than there was for the nominal mission, and we have to compromise somewhere. Returning to Earth time was just a side benefit. We're doing the best we can with the money available, but this just seems like such a waste.[1])

Compounding the problem, most of the scientists will soon be returning to their home institutions, leaving us to do more with less. During a lull, Justin Maki talks with some of us about a replacement scheme for taking post-drive panoramas. Until now, we've always figured out exactly what we needed and built a sequence to do just that -- custom-built solutions make for efficient operations. Justin's idea is to generate a bunch of panorama sequences up front, and then they'd just choose one of the canned sequences. The canned sequences won't be exactly what you want -- you might have to take more images to cover the region of interest, for instance -- but we'll be able to do it faster, because we won't actually have to build the sequences each time. He's going to write a Perl script to generate the sequences automatically, so I try to talk him into providing the script to the scientists, so that they can generate custom panoramas quickly. This would be the ideal solution, as I see it, because we'd still be able to build the sequences fast and yet we'd be making efficient use of the rovers. But I'm unable to convince him, and eventually I give up. Inefficiency wins again.

As it happens, today's one of the days when our current level of inefficiency bites us. I've been chewing on the MTES problem the whole day, and on my way out, I think of the solution. The problem was that the rock we were aiming for was off to one side, so even a small amount of slip could cause us to miss it. We should have driven straight at the rock instead; this would have mitigated the slip effect, making it more likely that we'd at least get a partial result. If we'd been on the old schedule, we'd have thought of this in time to do it that way. But it's too late now; the commands have already been uplinked. At this point, we'll just have to trust the rover.




[1] And, to be fair, practically everyone but me hated working on Mars time -- at best, they were exhausted and were quite relieved to switch to an Earth-time schedule. As inefficient as the Earth-time schedule might be, it's even more inefficient to burn out your best people so that they leave the project for another one that lets them see their kids. I get it. But if I were the only person concerned, we'd be on Mars time to this day.

No comments: