Opportunity Sol 875 (Spirit Sol 896)

So close, oh, so close to Beagle. Less than a hundred meters now. But Mars isn't making it easy for us. Around Beagle, there seems to be an earthen (marsen?) rampart, a huge wall of impenetrable ripples. As far as we can see, north to south, the darn thing is just too tall to get over.

And if we just got over this one ripple, we'd have pretty smooth sailing all the way. On the other side lies the Beagle Highway, a sizable stretch of outcrop we could follow all the way in.

Yestersol, they tried the obvious approach: just power over this damn thing. It was risky, but they were careful -- still smarting from our recent experience in Jammerbugt, very careful -- and therefore sequenced quite conservatively.

And it was a good thing they did. We ended up getting partially buried in the ripple, almost as soon as we got fully onto it. But, because of the previous RPs' conservatism, we're not so badly buried that we can't simply back downslope.

Which is our plan for the day: just back away slowly. It's a short drive, but not a short day: as someone points out, discussion = 1/distance.

Before we go too far down, though, we're going to take advantage of the fact that we're about as high up as we're going to be for a bit: we'll snap some pictures of possible alternate routes to Beagle from this relatively high vantage point. We hope they show us something good, because we haven't got a whole lot of promising routes at this point.

By which I mean, obviously: we haven't got any. We're hoping the images will show us at least one.

Of course, there's always this option: just bypass Beagle altogether, get onto the sand sheet where we can move freely, then loop around and come at Beagle from the south instead. Rampart, schmampart: treat the giant ripple as a Maginot Line.

After we're pretty much done for the day, our old pal Jim Erickson stops by for a provocative discussion. "What tools would you want to have to drive the next generation of rovers?" he asks. I think he's just giving me an opportunity to plump for using RSVP on MSL, but he's after more than that.

So far, he points out, we've basically kept single drives within a known range of terrain types. When terrain types change significantly -- e.g., when Opportunity went from flat plains to the ripple terrain -- we've been able to adapt slowly, even to the point of uplinking new flight software to deal with the terrain changes. What he wants to know now is, how can we go beyond that? How would we have to go about driving a rover if a single drive might carry it across multiple terrain types, some of which we hadn't seen yet?

Hmmm ... now that's thinking big. I think we'd want some way to recognize and categorize terrain types, then have the rover switch to different strategies based on what type of terrain it thought it was in. We'd have a collection of strategies for known terrain types; for truly novel terrain types, it might have to go around or wait for help.

That implies a lot of changes to RSVP. Time to get to work.

[Next post: sol 908, July 23.]

No comments: