Saturday, March 31, 2012

I'm gonna go with… Mars Quake. Because I can.

(Alternate title: "I have eighteen theories regarding the Opportunity phantom movement anomaly, and unfortunately none of them include aliens...")

Sometime between sols 2894 and 2899, the rover moved. Which is cool and all, since that's what a rover is supposed to do. Except when it's not supposed to.

Subtle; notice that it was only the left wheel:

Courtesy user elakdawalla from UMSF -- thanks!

This is significant mostly because it's a mystery why it happened at all. It is not significant from a rover health perspective. Put a two dozen engineers in a room and tell them that only some nearly indeterminably small thing happened, and you'll suddenly have two dozen engineers picking each other's brains for hours. The thing happened, the thing wasn't dangerous… but my god, that thing has an explanation somewhere. It bugs us when we can't explain something.

Of course, when you're not expecting anomalies with the rover, you're not taking a whole lot of pictures and you're not collecting a whole lot of data. We certainly didn't expect any kind of movement and therefore the number of pictures at our disposal to catalog the sequence of events is rather low. 

We know a few things: 

> It occurred sometime over a range of a few sols
> It occurred around IDD activity close to the surface
> It is possible that the right rear wheel moved as well
> It is possible that there was additional movement on sol 2900 or 2901

We don't know some other things:

> Our exact attitude (orientation) before the movement — more on this in a bit
> When in the possible sequence of events the movement happened
> Why the movement happened

Now, not only do we not expect the movement of the rover at all, we don't want it. Although the change in orientation is on the order of a few hundredths of a degree, and although it doesn't affects the amount of sunlight hitting the arrays because the change in orientation is below our level to detect the corresponding change in solar power reaching the rover, and although we can easily adjust for such a small movement for high-precision IDD work, there is one thing for which this matters:

Radiooooo Ssssscience

(Say each word with a pause, imagining that you're both a) Neil deGrasse Tyson and b) in a big room. Say it out loud, too, no matter where you are.)

Alas! Radio Science! 

The collection of the radio science data is pretty much invisible to the amateur MER followers. It doesn't produce pictures, it doesn't move the rover, and it doesn't have that wow factor… Well, at least at first glance. To the MER team, radio science has been this thorn in our side because Opportunity simply didn't have the power levels to execute as many as we wanted. We wanted all those radio science tracks to get in, but it's just so damned hard. Sometimes you win, sometimes Mars wins.

A few weeks back, the guy in charge of the radio science campaign gave a presentation to the MER tactical team regarding the data he had collected so far. Simply stunning. The science all boils down to measuring the doppler shift*, subtracting out the components of this shift that are due to known sources, and seeing what's left. What's left is the signature of the planet's wobble. Right, so, why does such a small movement of the rover — on the order of millimeters, as are as we can tell — matter?

The signature of the Mars' wobble, like that of Earth, is very small. Very, very small. The changes it induces in the shift of the radio signal are on the order of… well, something really small. When that is mapped into a velocity, it's under a millimeter per second. The exact position of the rover doesn't matter for the collection of the radio science data, as long as it doesn't move; if it does move, the movements need to be known down to below the centimeter level of precision. (We can measure that, by the way, using the rover's VisOdom software; we just chose not to move due to power levels this winter. Two birds, one stone.) If the rover moves, and we don't know exactly how or where, this has a magnified effect on the quality of the radio science data.

Right. So. We shouldn't move.

I think I can speak for the MER team when I say that our current best theory for the movement is that this wheel is perched on some crumbly rocks. All the math and mechanics and discussion we throw at it boil down to that: just crumbly rocks. It's a decent theory and it explains most of what we've seen. (Ahem… most.) 

In the coming sols, we'll be taking lots of orientation data and lots of new pictures to get a high-fidelity feel for our current state, so if it happens again, we have something against which we can take a difference. We've got roughly a month left until we will start moving again — we have to wait for the sun to get higher in the sky** and we want to collect another months' worth of radio science to get those error bars down.



*Yes, I'm still working on that promised Radio Science post. It will include all activity up to now, a brief overview of the basics of the science, and how we chose to squeeze it into Oppy's daily activities. Big deal!
**Happy Solstice! 03/31/12, winter solstice hit the southern hemisphere of Mars. Recall that minimum solar insolation was a few weeks before that.

No comments: