Scott Maxwell woke from uneasy dreams one morning to find himself changed into a Mars rover driver.
First thing out of the elevator I run into John Grant and Bill Dias. "We drove," John says. "You knew that, right?" I didn't. Fantastic as it is, that still might not be good enough. Maybe we only did the backup, maybe we didn't test Mark's autonav -- the main point of this drive. "How far did we get?" I ask him.
"I don't know how far we got, but Adirondack's not in our field of view any more."
I race to put my stuff away and get to the SMSA, where they'll have the answers. Adirondack not in our field of view would mean we did most of the drive. We got past the backup, and drove at least far enough to pass the rock we've been staring at for half a month.
The SMSA is, collectively, in its happy place. Nobody's happier than Mark: the autonav worked. ("One small step for MER, one giant leap for autonomy," Eddie Tunstel emails later.) And just exactly the way we wanted it to -- we got all the way to the end, we didn't time out. Spirit took her first three steps solo. There's a low-resolution thumbnail HAZCAM movie showing our traverse -- seven images, 40 cm apart. You get to watch Adirondack pass right under us. Soon we'll get the full-sized images.
The total distance we drove nearly tied Sojourner's one-day record. Tomorrow we won't just break that record, we'll smash it. Today we doubled our total mileage; tomorrow we'll do that again, and then some. At last, we're on the move!
There's some question of whether we clipped Adirondack during the drive. The original simulations had us missing it completely, but for some reason we clipped it slightly (in the simulations) after the backup, even though we returned to the same point before starting the body of the drive. There wasn't any threat to the rover, but I was hoping to avoid stepping on Adirondack, just on principle. The images aren't quite good enough to tell us, and I get distracted by other things before I can check the playback data.
Part of what distracts me is Ray Arvidson, who's already getting down to business. Tomorrow is going to be a different kind of sol -- all driving, very little science. Our tentative goal is to do 15m of fast blind driving, then 15m or more of autonav. We might not be able to get started until late, though, because the NAVCAM images we need for planning won't come down for a few hours. Fortunately -- considering that I got practically no sleep -- Chris Leger is the primary rover driver today; I'm just going to be his shadow. But between us, we'll have a lot to do, and we might not have much time to do it in. We'll need to have resource usage estimates by the downlink assessment meeting at 1600, less than thirty minutes away. He asks if I can work out a timing estimate for the drive, based on yesterday's behavior, which I do -- but only after Craig Leff taps me on the shoulder to show me something. It's a freshly received 3-D rear HAZCAM image, clearly showing Adirondack behind us, with the tracks leading by (or over?). "Happiness is seeing Adirondack in your rear-view mirror," Craig says, and I couldn't agree more.
I did it. I did it!
Courtesy NASA/JPL-Caltech. "Happiness is seeing Adirondack in your rear-view mirror." -- Craig Leff. I couldn't have said it better myself.
At the downlink assessment meeting I'm so excited I literally can't sit down. I pace the floor the whole time, feeling like a caged tiger. Jim Bell walks by one of the projection monitors next to me, which is displaying the rear HAZCAM of our drive tracks. "That's awesome," he grins, gazing up at it. "Great driving!"
Adler gives the engineering report, which is basically: it worked! And you guys won't be doing much tomorrow, because we're driving. Everybody seems OK with this, which is mildly surprising, but only mildly. Andy also speaks briefly, reporting the problem that caused us to lose half of the APXS integration the other night, and asking for the PULs to communicate the timing of their sequences upstream. Everybody seems OK with this, too. We're one big happy family.
The long-term plan is to pick destinations 80m apart and try to get there in 30m steps, one step per day, taking forward and rearward NAVCAMs after each step for localization. We'll stop for science at the 80m destinations, but not at each step along the way.
Arvidson is sitting in the back of the room with Chris and me. He leans over and puts in a request: at some point, he'll want to use rocker/bogie angles to reconstruct the terrain we're passing over. This is the first drive toward a crater on another planet. He mentions that it should get bumpier as we go, which makes perfect sense but I never thought about it before.
Yestersol, the tentative plan for today (thisol?) was to swing by the lander on the way to Bonneville, so that we could image the astrobot -- a little LEGO man, surrounded by magnets, placed on the lander by the Planetary Society. The magnets are yet another scientific experiment for measuring the magnetic properties of the Martian dust, and the LEGO man -- well, he's a LEGO man, damn it. Nobody seems to be mentioning this drive-by in the meeting, so Walter Goetz and I remind them about it. I don't know how important the science is, but I grew up playing with LEGOs, and I'll be damned if I'm going to let this chance slip by us! So he and I subtly talk the group into giving our little astrobot his fifteen pixels of fame.
At some point in the discussion, somebody describes the astrobot as a person ("a little figure of a person," was the exact phrase, I think), and this draws a rebuke: we are not to refer to the astrobot as a person, a human, an astronaut, etc. He -- it -- is a caricature of a robotic explorer of Mars, that's all. Incredibly, this was a decision made at the White House level, way back when the astrobot was first being considered for inclusion. At the time, the White House policy was firmly that there were no plans for a manned mission to Mars, and because of this we were absolutely not to characterize the astrobot as human in any way. This policy has not changed as far as we know, despite the Bush administration's call for a manned Mars mission, so the rule stands. The term "astrobot" is officially OK, cleared by the White House or whatever.
Okay, then.
That excitement ends with a decision to image the little figure of a person as we drive by, just as discussed yesterday. It's not a sufficiently important observation to justify a go/no-go cycle, as that would eat significantly into the main purpose of the day -- driving. So we're going to drive by, take our best shot, and keep going without a human decision cycle.
Courtesy NASA/JPL-Caltech. More information available from The Planetary Society and JPL.
On my way out, Jack Farmer, one of the LTP leads, congratulates me on the drive. When the images came up -- shortly before I arrived -- people were cheering on the VOCA net, he says. Damn, I wish I'd been there.
Chris and Mark Maimone and I sit down to plan out a tentative drive. ("Don't hit the lander," Mark offers helpfully.) It's actually pretty straightforward. Our rough plan is to drive past the lander, stopping to image the astrobot, then continuing, for a total straight drive of about 13m. Then we turn towards Bonneville -- a counterclockwise turn of about 45 degrees -- and turn the autonav code loose.
That's not the problem. The problem is figuring out what resources we'll need to get this done, so that the scientists can dispose of the rest. This leads to a scramble, as we try to find the relevant documentation and figure out how to enter the data into SAP, which we don't normally use.
What's more, Mark Maimone wants us to gather a huge amount of engineering data, 52 Mbits or so of images that will document the drive and enable him to tune his code here on the ground. This is going to eat significantly into the science allocation, so we expect trouble at the SOWG. I grab him and bring him upstairs with us.
This turns out to be completely unnecessary. The scientists don't complain at all. We get everything we need.
Sequencing the drive is not much of a challenge. Chris has it mostly ready in no time. It's complicated a little by the fact that we have to command no more than 1m of driving at a time for the blind driving, with stops of 50 seconds in between, to ensure the rover's actuators don't overheat. This is a new rule we just found out about today. But that doesn't make things much harder -- once Chris gets the two straight-line drive segments planned out, he goes back and breaks them up into chunks. This also gives us a natural way to perform a crude sort of course correction -- after every few meters, we aim the rover at its destination again, just in case it yaws while driving over a rock or loose sand or something.
It's also complicated by the behavior of the autonav code. We want to see how far the rover will drive on its own, but it has to stop by a certain time so that we can communicate with Earth in the afternoon. Our solution is to set a mobility time-of-day limit to stop the drive about an hour early, autonav until the limit, then clear the time-of-day error and take another short step forward (with documentation imagery) so that we can be sure we'll know what's under the IDD in case we decide to deploy it tomorrow. We have to fool around with it because of some last-minute problems, some with the planning and one that's a stupid and worrisome bug in RoSE, but it all comes together in the end.
On my way out, I walk through the SMSA, where Mark Maimone is looking at the NAVCAM images from the day's drive. He flips through the side looks that were taken during the drive, and at the backward-looking image of the rover's tracks. "Why would I ever go home?" he asks.
I also run into Richard Kornfeld, who tells me that, sadly, the Swiss TV producer cut the bit where I was on screen. But they did mention that I was filming, so I'm officially a Swiss TV cameraman now. Any day you can say that is a good day.
No comments:
Post a Comment