2009-01-18

Spirit Sol 15

I arrive early but immediately get sucked into fixing a bug in someone else's software. This annoys the heck out of me and eats into the start of my day. Worst of all, dealing with it makes me late for the downlink assessment meeting. During the part of it I'm in, I'm too irritated and distracted to really focus on it anyway, so I'm a little relieved when Rick Welch pulls me out of it. "We're driving today," he says, "did you know that?"

We're driving today.

When every day costs four million dollars, you make the most of them. Eric's sequences are all going well, so they've decided to drive to Adirondack this afternoon, a day earlier than we'd thought. I start running down my checklist. Mobility/IDD has nothing new; next thing is to check in with ACS[1]. Steve Collins is a sharp guy I've almost known for a long time (don't ask me to explain that), and he's the ACS rep today. I check in with him. The HGA[2] is going to point at Earth through the PMA[3], so we might get some degradation in the post-drive DTE[4] pass. It's trivial to fix; we just have to tell the HGA to flip over to the other side. But they actually want to keep it like this, so that we can characterize the behavior of the HGA in this configuration. OK by me.

I am almost but not quite late for the SOWG meeting, arriving just in time for them to say, "Oh, there he is now." The SOWG is frantic, but I'm in the zone. Maybe I can do this job.

Mark Maimone has a few tweaks to Jeff's drive sequence, which he and I take care of during the SOWG, but I can't do much to it beyond that for now. I know Jeff wants to add some imaging to the drive, but I don't know what he wants, so I decide not to mess with the sequence until he arrives. Instead, I start reviewing Eric's IDD sequence. Some of the pointing looks wrong -- the values differ from my own calculations only in the third or fourth decimal place, and that won't change where the arm is placed in practice, but there shouldn't be any difference at all. Anything unexplained is bad. I fret about it until Bob Bonitz arrives, and we manage to shrink the differences a little further, but we never do figure out what's wrong. We need to add some other stuff anyway, to get a picture of the flag on the RAT, so we move on, unsatisfied.

When Jeff Biesiadecki arrives, the team splits up: Bob and Brian work on the IDD stuff, and Jeff and I take the mobility stuff. I don't really understand what he wants to add, and this is my chance to learn about it so that I can get it right if I have to do it myself. The change to the mobility sequence is to use the NAVCAM to look to the right periodically as the rover drives, so that Mark Maimone can tune his visual odometry code. There are some other changes, but none of them really throws me.

The complicated part, as it turns out, is something else altogether. The rover can't move during the afternoon DTE session, because the antenna bounces around too much to maintain lock, so we have to finish driving by then. But we can't start early, because the RAT flag image has to happen after 11AM in order for the light to be right. We can't start early, we can't finish late, and we can't cut anything out of the middle.

Or can we? We added a five- or ten-minute sequence requested by Mike Malin, one that will get an important PANCAM image of the disturbed soil near the lander. This will help tell us what the soil is made of and why it acts like it does, which is currently quite a mystery. So there's scientific merit in the sequence. We can't take the soil picture before driving because the rover's body is in the way, and after the drive we'll be too far away to get as good a picture as we need. We have to take it right in the middle. If we do it at all. But the drive is so tight already, and buying back five or ten minutes might make all the difference, allowing us to complete the drive and start science on Adirondack a day earlier. And I did warn them at the SOWG meeting that this image would be best-effort only.

Jeff and I debate this for a while and eventually decide to leave the imaging in. It's a risk, but it's worth it. Even if we take the imaging out, there's a good chance the drive won't complete anyway, in which case we get no science and we don't nail the drive. With the imaging in, at least we'll have something. We'll let the drive finish if it can, and if it can't, the master sequence for the day will have to simply kill off the driving sequence before the DTE pass.

So far there's nothing unusual about this arrangement; we must have done it dozens of times in testing. The catch has to do with the rover's Inertial Measurement Unit, which tracks the rover's attitude as it moves. The IMU draws power and generates heat, and today we're already low on power and too warm. So we don't want to leave the IMU on after the drive if we don't have to. But if the drive sequence is interrupted by the master sequence, the IMU won't get turned off, which would be bad for the rover.

We end up splitting the sequence into two parts, the second of which handles the cleanup.[5] Even if the driving sequence is terminated early, the master sequence will call the cleanup code, which will shut down the IMU. We also arrange to have the IMU shut down at the end of the drive sequence itself, so that if we finish earlier than expected, we'll have the IMU off for that much longer. Jeff and I are worried about the last-minute changes we had to make to rearrange everything, but we go over it and over it and can find nothing wrong. In a way, that bothers me -- surely we screwed something up. "What have we forgotten?" I ask, and we look again. But we don't find anything.

At this point I've been working, constantly stressed, for nearly three hours longer than my shift. I don't want to leave, but I know I need to. I've never been so exhilarausted. Jeff and I wish each other luck, and I make my way home. I go to sleep with my fingers crossed, at least metaphorically. While I sleep, a hundred million miles away, a robotic scientist is carrying out my commands, slowly turning and crawling across the surface of another world.

Or maybe not.



Footnotes:

[1] Attitude Control Subsystem.

[2] High-Gain Antenna. It's the one that looks like a lollipop.

[3] PANCAM Mast Assembly. It's the rover's "neck and head" -- the camera mast that sticks up and has the PANCAMs, NAVCAMs, and MTES on top.

[4] Direct-To-Earth -- that is, a data pass where we send data directly to listening radio antennae on Earth, not indirectly through an orbiter.

[5] This immediately became standard practice, something we do to this day. And that's one of the great things about working on MER surface operations. You're faced with a tough problem, you invent a solution on the fly, and that solution becomes the standard practice for solving those problems (though often, as in this case, it's incrementally refined later). It makes me feel like I'm on the bridge of the Enterprise, like I'm Kirk inventing new battle tactics as the Klingons are attacking and then the tactic gets named after me and taught at the Academy.

Okay, so maybe it's not quite that dramatic. But it's awfully stimulating, all the same.

No comments: