2009-02-14

Spirit Sol 42

While I was sleeping, Art Thompson called a meeting to discuss the megadrive approach. Mark and Chris were awake and got the email, went to the meeting, and ended up changing the approach somewhat. They're writing scripts to autogenerate a large number of canned drive sequences, with blind-drive distances ranging from 0m to 25m at headings of 20 to 80 degrees (centered on Bonneville), and a family of autonav sequences that will take us in similar headings.

I'm a little put off by this at first -- not just because I had already decided inside of my own head that this side project was mine, but also because it's starting to look like Chris will get to do all the long drives, and I'll mostly get wimpy drives like the one we executed today (all of 20cm, not counting the turn-in-place -- bringing our total odometry to 59.5m). But it doesn't take me long to get over it. I don't have time to produce the megadrive building blocks anyway, so it's a good thing Chris and Mark can do it. And getting to drive the rover at all is cool. I don't need to drive 50m. 1cm would be cool, and I get to do longer drives than that every day.

I repeat this to myself until I start to believe it.

I can't do any useful work because we didn't get all of the post-drive HAZCAM images down yet, so I catch up on my email, which has stacked up terribly. I'm losing that battle. I take care of the worst of it, then go to the downlink assessment meeting.

Happily, there's a lot of praise for me and Bob, as we did the most complex sequencing to date last night. Jennifer Trosper, reporting to the scientists on yestersol's activities, says that we did a "fantastic job." The SOWG chair, John Grant, agrees: "I know yestersol's plan was really complex, and we're amazed at the work you guys are doing to put it all together." Jennifer repeats, "Yeah, the overnight team did a fantastic job. We were surprised it worked -- as we usually are."

They're a bit stymied by the lack of HAZCAM data as well, but they know roughly what they want to do. Mimi was the main target, and they might be interested in some soil targets as well. Mimi is that flaky rock that was off our front wheel until we turned to face it yestersol. It's not a terribly large rock, just 15cm wide, and the PANCAM shows that it's not spectrally unusual. But they're still interested in it as a possible sedimentary rock.

While we're waiting on the data we'll need for more complex planning, they show one of the MI results from yestersol. I'm not sure whether we're looking at the dune crest or the trough in this image, but whatever it is, it's full of tiny little round grains or pebbles. Smaller than Meridiani Marsberries, and with a different composition, but not entirely dissimilar in appearance. Nobody understands them yet.




Courtesy NASA/JPL-Caltech. Sol-41 microscopic image of the dune, showing it's covered with small round grains or pebbles.



John leads a short discussion about these results, and I observe something funny about his style. He'll pose a question, then walk over to his interlocutor to hand them the microphone -- and in form, it looks almost exactly like Donahue. Also like Donahue, John often is kind of tightly wound, but to his credit, I must say that he plainly recognizes this and tries to overcome it. And he never takes it out on the engineers, as far as I can tell -- on the contrary, with us he's always genuinely appreciative and complimentary.

While they're talking about this, we get the news that the front HAZCAM image is in, so we'll be able to IDD tomorrow. The plan, as yesterday, is to spend one day IDDing here, then move on. This is going to be even more complex than last night, as we're going to do several overnight observations leading to a morning touch-and-go.

In order to do any of this work, I need a terrain mesh. Because of the delay in getting our downlink data, I decide to head upstairs and ensure that MIPL is aware that they need to produce one for us. I haven't interacted much with MIPL before, so I need to spend a long time talking to them about their process before I can figure out (a) what I need, and (b) what I should ask for. I also need to talk to them about why the reachability maps they've produced seem to show that we can't place the IDD on the top of Mimi, even though it's obviously accessible. The scientists work from these reachability maps when deciding what instrument placements to request, and I'm worried that they're going to miss out on doing the science they want to do because of the misleading data.

We work out those problems, but it takes time for the software to chew through the data and produce new results. And you never know exactly what's going to come out -- if there's a problem, and you get the wrong result, they have to start over, waiting another 15 to 90 minutes (depending on the mesh complexity). I head to the SOWG with my fingers crossed. (The first mesh they produce, it turns out later, is a mess, but Bob and Payam work hard on it for me, and we end up with a really nice product that's exactly what I need. And I learn a lot in the process.)

The headline at the SOWG is a rumor that the free ice cream will be terminated at the end of February. It's not clear whether money is the reason (much less that the rumor is true in the first place), but they're apparently spending $3000 per week on the ice cream.

That's a lot of ice cream.

While MIPL chews on the terrain mesh, we chew on tomorrow's plan. As expected, it's the most complex one we've ever attempted: a series of five sequences placing each of three instruments on two targets, including three overnight tool changes (two of which require careful attention to heating). One of the targets is Mimi, the other is our own wheel tracks, which are not normally in the IDD work volume (but they are this time because of yestersol's turn-in-place) and which will tell us something about the physical properties of the soil. I race to implement all of this before John shows up as RP-2, and do a reasonable job of it. As usual, though, there are a lot of niggling little details that require attention, and though I work on it well past the end of my shift, eventually I hand it over to him. It's in decent shape by then.

Even so, when I do leave, I feel as though I'm leaving John in the lurch. On the way home, I reflect on some ways I can do my job better. I'm a perfectionist, and that's really working against me here. I can't stand to give something I've been working on to someone else unless it's perfect; I hate saying, "Well, this is broken, and now it's your problem." But that leads me to delay handing off the sequences to the RP-2 (second-shift rover planner), which makes their jobs harder. My job isn't to get the sequences perfect, it's to get a decent first draft in place for the RP-2. Maybe if I think of it that way, it will neutralize my perfectionism. Also, maybe pigs will fly.

1 comment:

Anonymous said...

Well you can't just leave us hanging. Did you lose the ice cream?