The Design and Engineering of Curiosity

by Emily Lakdawalla

  3.3 STRATEGIC, SUPRATACTICAL, AND TACTICAL PLANNING

  Curiosity is far more capable than its predecessors, especially when it comes to drilling and sampling, but there are no more hours in the day for Curiosity operations to be planned than there were for Spirit and Opportunity. There are 10 science instruments, and more than 400 scientists, and planning has to be mindful of the needs of multi-sol campaigns. Performing daily operations with a complex rover while keeping eyes set on long-term goals is difficult for such an unwieldy team.

  Many different factors limit what the rover can accomplish in a given sol. Power is one: most activities draw power from the batteries at a faster rate than the MMRTG can replenish it, but for safety reasons the tactical team is usually required to leave the rover batteries nearly fully charged at the end of each sol’s plan. Communications are another serious bottleneck: the rover can capture far more data than there is capacity to transmit it to Earth. But at the beginning of the mission, the most stringent limit on Curiosity’s activities was imposed by the sheer complexity of the machine.

  To make it all work, Curiosity mission operations are planned on four different timescales, with all four working in parallel:The Project Science Group (PSG), a committee consisting of the project scientist (John Grotzinger at landing, and later Ashwin Vasavada), NASA Mars Program Scientist Michael Meyer, and the principal investigators of the science instruments, establishes the overarching scientific questions motivating Curiosity’s mission, and determines the very long-term driving destinations. The goals, questions, and destinations were established in the mission proposal and subsequent extended-mission proposals.

  The strategic or long-term planning process addresses development and testing of first-time activities, planning science and sampling campaigns, and long-term management of rover resources. Long-term planners map out a “sol path” covering several sols, a high-level list of activities that directs the mission toward accomplishing the Project Science Group’s goals. Strategic planning works on week- to months-long time scales, and is mostly conducted on Earth time.

  The plans needed to implement the strategic plans are developed first in a “supratactical” planning process. One to several sols ahead of time, the supratactical team sequences the “look-ahead plan”, beginning to sketch out the actual list of commands to be sent to the rover. At the beginning of the mission, the supratactical process was conducted on Mars time.

  Finally, the tactical team produces each sol’s plan. The supratactical team hands the tactical team an outline of the activity plan for that sol, called a skeleton plan, along with guidelines for what the rover needs to accomplish to stay on the look-ahead plan. Each skeleton plan contains science blocks, periods of time during which the tactical team can add in science observations, resource limitations permitting. The tactical team responds to data downlinked from the rover each sol, fleshes out the plan handed to them by the supratactical team, and generates commands to uplink to the rover for the next sol. Tactical planning was conducted on Mars time when the mission began, with the tactical planning timeline taking 1 sol, operating every day of the Earth week, around the clock.6

  Curiosity differs from its predecessors Spirit and Opportunity in having the supratactical planning process, which is necessary because of Curiosity’s complexity and the intensive resource demands of its analytical instruments, CheMin and SAM. In parallel with strategic planning, the supratactical process takes care of the negotiations among different instruments for rover resources, assigning activities to different sols to balance out demands.

  3.4 TACTICAL PLANNING PROCESS

  3.4.1 Mars time operations

  Since the rover needs a plan for each sol that responds to what happened the previous sol, the ideal way to operate Curiosity is to begin planning the next sol at the end of each active sol on Mars. When the mission operated on Mars time, the planners worked over the rover’s night to deliver a new tactical plan at around 10:00 local time each rover morning.

  Late in the afternoon (between 3:00 and 6:00 p.m. local time), both Mars Odyssey and Mars Reconnaissance Orbiter fly over the landing site. (Read more about rover telecommunications in section 4.​5.) Colloquially, a communications session is referred to as a “pass”, because the orbiter is passing across the rover’s sky as the rover uplinks its data. The orbiters relay the data onward to Deep Space Network radio dishes on Earth. The last orbiter communications session before Earth planning begins is called the “decisional data pass” because it is the last pass containing data that Earth planners can use to make decisions about rover activities. Decisional data includes telemetry on the health and safety of the rover, and Hazcam and Navcam images that can be used to build a terrain mesh, a 3D map of the terrain around the rover.

  The rest of the data comes down in priority order. The tactical team carefully assigns priority to every data product that they command the rover to produce. They assign highest priority to science data that is beneficial for planning – such as Mastcam images of the area that the arm instruments can reach, or ChemCam can zap; these usually arrive quickly. Other data may come down days later. The Mastcams, in particular, generate huge volumes of data and can store it for months inside large flash memory drives within each instrument’s electronics. Low-priority Mastcam data can easily sit on the rover for a year before downlink. The mission always tries to keep some volume of data in memory so that if an onboard anomaly prevents new activities, all available downlink sessions can be used to downlink science data. The team makes all these priority assignments during tactical planning, but can also reprioritize data still on the rover to force it to return to Earth sooner or later, as appropriate.

  On Earth, a downlink team of instrument scientists and rover engineers studies the downlinked data to assess rover health and suggest a set of activities. Responsibility shifts to an uplink team, which includes representatives of every science instrument as well as rover planners (also known as rover drivers). The uplink team generates a command sequence and sends it to the rover. Usually the Deep Space Network transmits the sequence directly to the rover’s steerable high-gain antenna around 10:00 a.m. local solar time, after the Sun has warmed the rover slightly, in time for it to begin its next sol of operations on Mars.

  Operating on Mars time presents two main challenges. One: the full cycle from data downlink to sequence uplink has to be completed within about 16 hours, over the rover’s night; if the sequence doesn’t get prepared by the end of that time, they miss their uplink window and lose a whole sol of activity on Mars. Two: Mars time and Earth time are not the same. To work on Mars time is to begin the planning day 39 minutes later each day. In 38 Earth days, there are 37 Mars sols. If the Earth and Mars schedules are perfectly aligned one day, then, 19 days later, the two schedules are perfectly out of sync, and operating on Mars time requires working through the Earth night.

  For the first 90 sols after landing, the mission operated on Mars time, through nights and weekends, with the whole science team co-located with the engineers at the Jet Propulsion Laboratory (JPL) in Pasadena, California. Mars time helped the engineers maintain a tight connection with the rover as they commissioned all its instruments and tools, and permitted them to use all of the 16-hour rover night to prepare each sol’s worth of activities. But Mars time is grueling for humans, whose circadian rhythms and private lives still run on Earth time.

  3.4.2 Slide sols, restricted sols, and solidays

  After sol 90, the science team members returned to their home institutions and the mission transitioned to Earth time. They had increased planning efficiency to the point that one sol’s worth of activities could be developed in a 9-hour planning day. Earth time operations are permitted to take place between the hours of approximately 6:00 a.m. to 10:00 p.m. California time. So in a day where uplink needed to take place no later than 4 p.m. California time, they could begin the planning day at 7 a.m., an “early slide sol.” In this way the team could operate as though they were o
n Mars time for about half of the sols, as long as they could fit a 9-hour planning period in work-permissible hours within the 16-hour window between the receipt of end-of-sol data from Mars and the time of the next sol’s uplink.

  As days and sols turn over, the beginning of Curiosity’s Mars sol creeps later and later in the tactical team’s Earth day. If decisional data arrives on the team’s computers after the planned start of work, they don’t have time to analyze the data from Mars before they need to plan the next day’s activities. For a couple of sols, they can slide the Earth planning timeline a little later in the day – starting, say, at 11:00 in the California morning and finishing at 8:00 in the evening – but these “late slide sols” only buy a couple of days on Mars-like time.

  An Earth planning day comes when the decisional data arrive from Mars too late for an Earth time schedule, in the middle of the day in California. On these “restricted sols,” the team has to plan the rover’s activities without any knowledge of whether the previous sol’s activities executed successfully. If the previous sol included arm work or driving, further motion is usually precluded until the planners can assess the success of the activities, so restricted sols following arm work or driving are heavy with remote sensing. (The engineers do consider it safe to move the mast head to perform imaging and ChemCam operations despite not knowing the state of the rover.) If the previous sol included a drive, the team has no way of knowing what the immediate landscape looks like around them. Therefore, they can’t conduct science work with the arm or target remote sensing. When the mission is in restricted sols, drives can only be commanded at most every other day. The mission’s first restricted sol was sol 92.

  Common restricted-sol activities include untargeted remote sensing in which cameras or spectrometers shoot in the blind. Some remote sensing observations don’t actually require detailed position information, like 360° panoramas, or imaging of distant targets whose positions don’t change much with one drive, like targets on the crater rim, the Gale central mountain, or sky objects like the Sun, clouds, Mars’ moons, stars, and comets. The team may also use restricted sols for SAM or CheMin analyses of samples already inside the rover. Sometimes restricted sols include little activity at all, an opportunity to let the batteries recharge. The restricted-sol period continues until the Mars clock drifts far enough with respect to the Earth clock for the 9-hour planning cycle to fall within the 16-hour window again. There is a “soliday” – an Earth day in which there is no need to plan for Mars, one every 38 Earth days – and then the mission comes in for two or three days of early slide sols to begin a couple of weeks of unrestricted planning. Whenever possible, the mission tries to plan solidays to fall on weekends.

  3.4.3 Weekends, holidays, and surge sols

  Even after the transition to Earth time, daily Mars operations imposed difficult demands on the lives of mission personnel. Operating through weekends was especially hard on workers with families. Knowing that the mission could continue for years, Curiosity management worked to reduce the mission planning schedule.

  As of sol 180, the mission ended routine Sunday tactical planning, planning two sols every Saturday instead. During restricted sols, the rover could be commanded to drive on Saturdays, to allow Mondays and Wednesdays to be used for planning driving sols, with Tuesdays and Thursdays used for untargeted remote sensing. Fridays could then be given over to arm activities and/or targeted operations with remote sensing instruments. Sundays were often used for time- and power-intensive SAM and CheMin analyses and/or untargeted weather observations.

  As of sol 270, the mission ended routine Saturday planning. From then on, Friday tactical planning covered three sols, or two with a soliday. And very rarely, the mission reactivated Saturday planning in order to take advantage of unrestricted sols for driving or drilling, but the project ended this practice in May 2015.

  For a few months beginning sol 515 and again on sol 635, in order to maximize the use of unrestricted sols, the mission employed the concept of “surge sols”. These are engineer-only (no formal science activity) sols with only a 6-hour planning period. They allow the team to begin planning very early or late in the Earth day and eke out another day or two of unrestricted drive sols before flipping over into restricted-sol operation. The project also performed surge-sol planning on weekends during unrestricted-sol periods, again without formal science team participation.

  After the end of the prime mission, around sol 765, the mission reduced planning days further, producing two-sol plans only three weekdays a week when in restricted sols. Nowadays, a typical 38-day/37-sol period begins with two or three early slide sols, then two or three weeks of 5-day-a-week unrestricted sol-planning, followed by two or three late slide sols, then two or three weeks of 3-day-a-week restricted-sol planning, then a soliday. At the start of Earth time planning, operations were more often in restricted than in unrestricted sols. As the team has become more experienced and the planning period has been shortened (to 8 hours), the mission now enjoys slightly more days in unrestricted sols.

  The mission has always reduced the intensity of planning during major United States holidays like Thanksgiving, Christmas/New Year, and Independence Day. They prepare multi-sol plans to tide the rover through these periods, usually focusing on routine environmental observations. Holiday plans don’t usually generate as much data as regular plans, so routine orbiter communications (which the rover handles autonomously according to a schedule delivered months in advance) during holidays are periods of catching up on data downlink.

  3.5 MISSION SUMMARY

  Recounting the daily operations of the rover is beyond the scope of this book. The following broad overview is intended to provide context for the discussion of how the rover’s systems and instruments work in the rest of this book.7 Appendix 1 contains a list of the official mission summaries of each sol of activity. A brief overview of mission activities is in Table 3.2.Table 3.2. Brief summary of major phases of the Curiosity mission.

  Sol

  Site/drive

  Date (UTC)

  Event

  0

  1/0008

  6 Aug 2012

  Landing

  21

  3/0100

  27 Aug 2012

  Drive toward Glenelg

  57

  5/0000

  3 Oct 2012

  Arrive at Rocknest

  102

  5/0388

  19 Nov 2012

  Depart Rocknest, drive toward Glenelg

  166

  6/0000

  23 Jan 2013

  Arrive at John Klein in Yellowknife Bay

  272

  6/0068

  12 May 2013

  Arrive at Cumberland, Yellowknife Bay

  324

  7/0000

  5 July 2013

  Depart Cumberland, begin Bradbury traverse

  392

  16/0050

  12 Sep 2013

  Arrive at Darwin (Waypoint 1)

  402

  16/0328

  23 Sep 2013

  Depart Darwin, continue Bradbury traverse

  439

  21/1572

  31 Oct 2013

  Arrive at Cooperstown (Waypoint 2)

  453

  22/0484

  14 Nov 2013

  Depart Cooperstown, continue Bradbury traverse

  535

  26/0366

  6 Feb 2014

  Cross Dingo Gap

  574

  30/0740

  18 Mar 2014

  Arrive at the Kimberley (KMS-9)

  634

  32/0204

  19 May 2014

  Depart the Kimberley

  753

  42/1020

  18 Sep 2014

  Arrive at Pahrump Hills

  923

  45/0558

  12 Mar 2015

  Depart Pahrump Hills

  992

  48/1194

  22 May 2015


  Arrival at Marias Pass

  1072

  49/0294

  12 Aug 2015

  Depart Marias Pass, travel south to cross the Stimson unit

  1172

  51/0592

  23 Nov 2015

  Approach Bagnold Dunes for first campaign

  1248

  52/0722

  9 Feb 2016

  Traveling west toward the Naukluft plateau

  1281

  53/1284

  14 Mar 2016

  Climb onto Naukluft plateau

  1369

  54/2508

  12 Jun 2016

  Turn south to cross the Bagnold dunes

  1427

  56/1326

  11 Aug 2016

  Approach Murray buttes

  1454

  57/2582

  8 Sep 2016

  Depart Murray buttes, drive south

  1508

  59/0936

  2 Nov 2016

  Enter southern Bagnold dunes

  1601

  60/3162

  6 Feb 2017

  Begin second Bagnold Dune campaign

  1671

  62/1140

  19 Apr 2017

  Exit dunes, traverse south to Vera Rubin Ridge (formerly known as Hematite Ridge)

  1726

  64/0000

  14 Jun 2017

  Arrive at Vera Rubin Ridge, turn east along ridge base

  1812

  66/0000

  11 Sep 2017

  Reach top of Vera Rubin Ridge