Saturday, November 12, 2011

On Mornings

Hi, my name is Matt Lenda, and I… this is very difficult for me to admit… and I am a morning person.

Hi, Matt.


Scott Maxwell's Twitter feed can be instructive:

"Turns out that if you start at oh-dark-frakking-thirty[*], you can finish waving the rover's arm around by jesus-please-tell-me-it's-naptime."

This is in reference to a rare early start to the day: the SOWG meeting's at 8am, not 9am. (Gives the Cornell folks a chance to tag up before lunch!) We had to pull this during our Nominal planning cycles this past week to squeeze in extra planning days. 

When I was in high school, I worked the open shift at a Starbucks. The day started at 4:30am; lights on by 5:30am. I didn't have real weekends during the school year for some two years. I averaged five hours of sleep a night — hockey practice, by the way, started at 10pm four nights a week — for nine months. I was able to add regimented naps in the summertime, but the schedule was pretty much the same: early in, early out. This turned me into a morning person, against my will as it were, and now I'm stuck in that groove. My body won't let me wake up later than 9am on the weekends unless I am up until 5am working on something; then it gets me up at 9:15am instead. 

When I was a kid I slept in until 10 or 11 on my off days just like anyone else. After working those early shifts a few years later, things had changed. I remember the first weekend I had had off in a few months: I got excited because I was going to get to sleep in for once, but my body wouldn't let me. It hasn't ever been the same since.

I am, characteristically, the first person into the room when I'm on shift. If I'm TAP/SIE, I clean my area. If I'm TUL (shadow, for now!), I clean the whole room. It's easier to admit that I have OCOS — Obsessive Compulsive Oooohhh Shiny! — than it is to admit that I'm a morning person. Somehow it seems to be the dirty word around the office. During my first week at JPL a year ago, I met another guy who had then himself only been at JPL for a year. He asked me when I planned on going into work. I answered, "7:30am, usually." He said, cynically, "Yeah, that'll change."

He didn't know the extent of my morning person-ness. They never did change, my arrival habits. In fact, getting there at 7am or earlier has been my recent thing. There are things to get done. There are spacebirds to fly. There are Mars cars to command. 12 hours later I wonder where the day went, thinking only, "Thank god I got here early."

Scott told me later that they used to allow 6:30am start times for the day. His, ermmm, distaste for such an early start aside, we have on our hands an interesting thought experiment. How do you balance the needs and desires of a tactical team? We have morning people and evening people. We have east coasters and west coasters. We have people who like light and people who like night. What's the best way to do this kind of thing?

The solution for a 3-month mission is easy: Shove everybody in the same location and live on Mars time. If Mars days were significantly longer or shorter than Earth days, that would very drastically change the design of the entire mission. Just as it is easy to forget that your most precious resource is time with the DSN, it is easy to forget that Opportunity literally lives and breathes by day. To say that such a thing as local time systems would entirely change a mission design is to say a very profound thing. The optimal solution is to move everybody to a single location and then adjust our finely-tuned body clocks to start cranking out sequences on Mars Time. To hell with Earth time — who says I can't have dinner at 4:33 am? MY BODY SAID IT WAS OK.

This is, by the by, our first solution: Live On Mars Time (read: "The Golden Days")

We can also call this "inconvenience everyone equally" paradigm. But, of course, Opportunity and Spirit lasted wayyyy longer than 3 months. The operational (and personal) costs of having the entire science and engineering team at one location are astronomical — but, of course, with huge payoffs if the duration of their stay is short — so the MER project had to find a new way of doing things. This is our second option:

Option 2: Modified Earth Time (read: "What we do right now you daft fool")
What do we mean by "modified"? We mean that we schedule tactical planning shifts within some convenient day time range; we can place constraints on what's allowed and what's not allowed. For instance, we can institute a rule that states, "No planning day should start before 8am," or equally, "No planning day can exceed 10 hours in length." Stuff like that.
The problem with this is that Mars and Earth rotate at speeds just different enough to be a total pain. There are many ways to show how, but the simplest is to plot local solar time in seconds for Mars and Earth — using Opportunity's longitude and the longitude of Pasadena, CA:

See how long it takes for the full cycle to go by? Total pain. We look at this drift of time well ahead of time, and when we see that the previous sol's downlink isn't going to get down to the ground in time before a reasonable start to the day, we "back off" a few days and plan multi-sol sets of commands for the rover. There are many rules and guidelines and formulae that go into determining this strategic schedule, but you get the drift. (*rimshot*)

The idiosyncrasies of this so-called "Modified Earth time" happened to hit us this week. We've been boogeying north to get to the where the north-facing slopes on Cape York happen to be. We had to get this terrain-mapping done before MSL launches because they will likely be "stealing" our uplink and downlink times with the DSN, meaning we can't command Opportunity as often as we'd like. MSL is, after all, a much higher priority spacecraft. By the time MSL stops taking our windows from us it might be too late to do a good reconnaissance of the surrounding areas. Power is dropping as we approach winter solstice; we can only support drives on the order of an hour in duration at the moment, and we can only do drives on two or three consecutive sols before we have to take a breather and let those batteries charge up.

What happened is that we banked ourselves some meters: the RPs got into "trek toward Endeavour" mode and just took off.** We had about a week of magian to get the terrain mapping done. Then: Homestake. And then: Deadwood. (More on these targets in later posts!) And then: all of our margin was gone. So the science folks gave us back the keys, saying, "Get 'er done." To do that, we had to institute a planning day today, Saturday. A weekend planning shift hasn't been necessary since, purportedly, previous Spirit winters. In that case, it was to literally save the spacecraft from certain doom. In this case, today, it was to put some more meters in the bank. Because we need ground in the loop between drive and IDD activities, a full 3-sol plan on Friday (the usual planning schedule) would have only given us one drive when we could have two or three. Damn Modified Earth Time! So, to counteract this, we came in again today to get some more driving time.

(Totally killer day. I felt like I was in the glory days, part of the prime mission, part of a team that was ecstatic to be driving cars on Mars on a Saturday morning.)

Really, everyone seems to miss the Glory Days of Mars Time. But I think everyone would be sick of being at JPL again, as nice as this place is. Which brings me to our third option for tactical operations:
Option 3: Floating. Tactical. Cruise ship. (read: "Mars Time for Winners")

Scott told me about this. The idea: Screw being spread across a continent, and screw being on land in a single stuffy building. Put everyone on a cruise ship. Travel west to make your day about 40 minutes longer than the usual Earth day.

40 minutes drift per day = 0.67 hours drift per day = 0.0278 days drift per day
0.0278 days drift per day * 360 degrees longitude per day = 10 degrees of longitude per day ~= 600 miles west every days at the equator

That's, like, a Colorado and a half every day. Totally feasible.***

Not only would this mean that we're on the Mars clock — ahem, Opportunity's clock — but we would also have the "normal" daylight hours where the sun rises in our mornings and sets in our evenings. Walking outside wouldn't totally mess with our fragile brains.


*"oh-dark-frakking thirty" is, by the by, an actual time, though one that's not institutionally accepted. We have to translate between this colloquial clock and UTC pretty regularly. I find it unfortunate that the rover's command language does not accept colloquial times as parameters: "Start APXS integration at cold-as-crap thirty-seven, wait for god-knows-how-long, shut your face, and salute me. Dang it."

**Days Is Almost Gone" by the Derek Trucks Band is currently blasting on my record player. It has grabbed my attention at this moment for two reasons: first, it seems apt for the pre-Endeavour attitude that was all about putting back the miles and having to shirk it this mindset when we hit Cape York; and second, it reminds me that my record player's needle isn't of very high quality and is making all of my 180-gram records hiss on the high ends, like vocals and drum cymbals. This is unnerving — having music of such high audio quality that it actually sounds worse. 

***I'm leaving the logistics of such an operations architecture as an exercise for the reader. Suckers. Also, I didn't check my math. Also, I have to go flip the record… Also, nobody is allowed to point it out if I screwed the math up. 


emptyshirt said...

If the joint ESA-NASA ExoMars mission comes together as hoped, there would likely be two mission controls, one in Europe, one at JPL, roughly 9 hours apart. Not as fun as a cruise ship, but addresses some of the Mars time issues -- it's always 5 o'clock on Mars somewhere! Margaritas anyone?

Kim said...

Too late for MSL (*sigh*), but 2018? Oh... wait. (but we could *totally* rig up a cruise ship for tactical operations by then! Complete with free ice cream! ;)