Practicing Mars 2020 rover operations, on Earth

February 4, 2019: My great-grandmother Schneider was reportedly fond of saying that “a birth, a death, and a marriage always happen close together.”  After years of planning, all the components on the Mars 2020 rover were now nearing completion, and soon the rover would be “born” and ready to stretch its legs in a clean room on Earth.  But in Endeavour Crater on Mars, 5000 km away from Jezero Crater where the Mars 2020 rover would touch down in two years, we had not heard from our old robotic colleague Opportunity since June when a thick dust storm blocked the sun and stole life-giving power from its solar panels.  The storm’s retreat left a thick coating of dust that not even the trusty, gusty winds on Mars could remove in time to save the rover.  After a 15-year mission that was supposed to last 90 days, it was looking more likely that its service had finally come to an end.  But we had a new rover to make ready.  And as part of the planning leading to launch in July 2020, the mission operations staff had begun the long process of marrying the science team into a cohesive unit, capable of operating the rover efficiently and with an eye toward capturing as many samples as possible from the red planet for eventual return to the earth.

A birth.  A death.  A marriage. I wager that Grandma Schneider never thought her bit of old wisdom would ever apply to Mars.

Months of preparations by the Mars 2020 mission operations staff would culminate in today’s first practice “sol” (martian day) as part of an exercise named the Rover Operations Activities for Science Team Training (ROASTT).  In this test, a team of scientists and engineers had secreted themselves away to an undisclosed location in the desert southwest, and began taking pictures and other data from a mobile platform that simulated the view provided by a rover.  Simultaneously, a team of scientists stationed around the world would meet remotely by telephone and computer video links to simulate nearly two weeks of rover activities, exercising all the software and communication tools needed to remotely operate the spacecraft.  Many of us had experience with operating the Mars Science Laboratory (MSL) Curiosity rover, currently exploring the base on Mt. Sharp in Gale Crater.  Some also had experience with the Mars Exploration Rover (MER) missions—Opportunity and its twin Spirit, which had lost its battle to the martian elements back in 2011.  Still others began their careers with Mars Pathfinder back in 1997 and its plucky little rover Sojourner.  A few stalwart gray-beards went as far back as the Viking missions of the late 1970s. 

Others of us had experienced these types of tests in preparation for the MER and MSL missions.  In 1999, NASA sponsored a Mars rover simulation in Silver Lake, California with a robotic prototype rover called Marsokhod.  In 2001, a similar field experiment was done with a rover called FIDO to practice for what would become the MER rover missions. I had been on both sides of the experiment: Sometimes I was in the field acting like a rover and taking data requested by a remote team hundreds of miles away.  Sometimes I sat comfortably in air-conditioned rooms, analyzing data that came in on a daily basis, interpreting what it meant and helping figure out what the “rover” should do the next day.  The goal was always to test the abilities of both the on-site and off-site teams to maximize the scientific return from a remotely-operated rover.  Like any good field trip, however, the ancillary goal was to get to know your colleagues—your team—and how best to combine sets of expertise while navigating the terrain and the rover. 

So what did we know about the area in which the ROASTT would be conducted?  We had been shown the location of the “rover” in orbital images and the intended path toward different points of interest in the rugged desert terrain.  We had been shown the “downlink” of data from “Sol 99” which revealed a narrow channel with rubble and small cobbles at the rover’s wheels, flanked by jumbled conglomerates of rock and dirt comprising the banks of an apparent channel eroded into the terrain.  In the distance, layered and tilted outcrops of rock were known to be the target of a 30-meter long drive that would happen on Sol 100.  A portion of the team was dedicated to the “Campaign Implementation” (CI) aspect—figuring out what strategies and priorities should dictate how and where the rover should conduct its business of exploration and discovery.  They had formulated a plan for Sol 101, the first real day of the test.  I was filling the role of the “Tactical Science Lead” (TSL), tasked with organizing members of the team responsible for crafting a rover activity plan for the sol that met the overall goals handed to us by the CI team.  

At 07:07 AM at the Jet Propulsion Laboratory in Pasadena, California, where the main mission operations hub was housed, engineers released the contents of the Sol 100 downlink to the team.  I had the luxury of living in Maryland where the corresponding 10:07 AM release time was a bit more manageable.  From the first look at the images, it was apparent that the drive on Sol 100 had been a success and we were less than 10 m away from “Waypoint 2”, a region identified from orbit that would provide access to the highest position in the area where exposed, layered outcrops of rock were thought to contain clay and carbonate minerals of interest to the science team—and prime targets for sampling! 

Our agenda for the day was highly scripted, with a series of timed meetings designed to advance the team from identifying rock targets of interest for analysis by the cameras and other instruments to piecing together “components” of activities that the rover and its instruments would perform.  For example—we could request that the rover take additional color photos or mosaics of part of the channel, or use the spectrometers to take multiple shots at specific locations on a rock outcrop to determine its chemical makeup.  Our tactical team had a little more than 3 hours dedicated to coming up with a plan that met today’s goals---survey the area up ahead well enough that the next sol’s short drive to Waypoint 2 would have enough information to begin interrogation of specific locations using the instruments on the rover’s arm.  That data would help the team pick a spot to drill, and eventually “extract” a sample to store onboard.

The “Kickoff” meeting began the day at 08:00 AM, with about 20 people with specific roles on shift and a hoard of other science team members on the telecon line to plan Sol 101.  In that meeting we confirmed that the rover was healthy and ready to proceed with another day of planning.  We learned about the other limitations we would have on planning, like the amount of data volume that would make it back to “Earth” before the next sol’s planning.  That would inform us about how to limit our appetites for new images and data, and how to prioritize what we acquired.  We reviewed the plan handed to us by the CI team, and made sure we read their reports about their intents for the observations in that plan.  Their plan included a large Mastcam-Z mosaic intended to image the channel wall on the left side of the rover and the area it would drive to next.  It also included three SuperCam observations to understand how the composition of the layered rocks near the rover might change from one place to another.

Over the subsequent 60 minutes the “Plan Development” meeting allowed us to discuss the scene around the rover in more detail and identify potential targets of interest.  The Mastcam-Z Payload Uplink Lead (PUL) for the day let us know that the intended 360-degree panorama in the CI plan required too much data volume (almost 10 times the amount we could downlink that day) and too long a time (four and a half hours) to be feasible.  So we began to construct a more tractable Mastcam-Z mosaic that would image the entire channel wall and the Waypoint 2 region but at lower resolution, cutting both the number of images (data volume) and time.  Others on the telecon pointed to the hill off the right side of the rover, and suggested that a lower priority observation would be a small mosaic of that region, given that it was the highest area in the local surroundings. The PUL began busily working on different efficient options to image the two areas. 

We heard from the SuperCam data analysts that the one observation acquired in the Sol 100 plan revealed the likely presence of clay minerals and possibly organic material.  This analysis was on a portion of bedrock that appeared similar to that exposed at the Waypoint 2 area.  Although Waypoint 2 rocks were enticing, they were about 7 meters ahead, which was just beyond the range that the SuperCam laser could be used for compositional analysis.  So the science team decided to use their allotment of three SuperCam targets on similar outcrops of rock that were closer.  This would help us understand any changes in composition using the combined efforts of Raman, visible/near-infrared, and laser-induced breakdown spectroscopy (LIBS).  Meanwhile, the Mars Environmental Dynamics Analyzer (MEDA) instrument would be on in the background during the sol, taking its periodic measurements of wind speed and direction, temperature, and humidity of the atmosphere. 

The next meeting was more of a formality to confirm the intent for the day and overall structure of the plan--the “Declare Plan” meeting. We already knew the guts of the plan—a couple Mastcam-Z mosaics, several SuperCam targets, and atmospheric monitoring.  So this meeting only lasted a few minutes, and led back into more detailed planning to finalize the details of the other observations and how they would be implemented in the plan.

Indeed, we spent the next 80 minutes on the “Plan Implementation and Validation” meeting where the order of how the observations would be made was analyzed.  Were any shadows creeping across our targets at the time of day they were planned?  If so, could we move the order of the activities to prevent that from happening?  What about adding a fourth SuperCam target on the Waypoint 2 area?  Even if the instrument’s laser couldn’t be used for LIBS and Raman observations, it could still acquire detailed measurements of the reflected sunlight off the rocks and mineral information that data would reveal.  Considering how much time we had saved by downsizing the original Mastcam-Z mosaic, it looked like that extra SuperCam target would fit without a problem.  We also re-evaluated the areas covered by each of the Mastcam-Z mosaics, confirmed that single Mastcam-Z images would be taken of each of the SuperCam targets for documentation, and even were able to accommodate a late request from the atmospheric scientists to include another use of MEDA to search for clouds in the morning of the next plan. 

An important part of our planning process involved writing down the “why” of our observations to help the next shift of team members understand what we were thinking.  What was the final intent of the observations?  As an example, the team worked on these pithy but effective summaries of the mosaics: 

• Channel mosaic: To characterize the layering and textures within channel deposits and outcrop exposures near the rover and document the Waypoint 2 area. 
• “Eastern Block” mosaic:  To image the dipping, light-toned beds and possible capping unit, layering, and textures within the eastern ridge rocks.

Our plan was close to being completed.  The PULs had done the majority of their work on the sequences, checking the final details and pointing.  Now the final meeting of the day--“Reconciliation”-- provided the team a last chance to check the plan, its data volume, and duration to make sure everything was within the limits we were provided.  It turned out that the tools we were using for the test were not yet up to speed on calculating all those things exactly, but we all agreed that our totals were within range of our limits.  After that was all buckled up, we declared victory in the tactical process, and dismissed the rest of the tactical team.

While our Tactical team had been preparing that sol’s plan, the CI team had been working at the same time in their own sets of meetings, preparing the Sol 102 plan outline.  My final meeting of the day (“Sol N+1”) would involve discussing their Sol 102 plan, which would involve some more remote sensing prior to the short drive to the Waypoint 2 location.  After the drive, more images of the “workspace” in front of the rover would be taken to help with selecting areas for investigation by the instruments on the rover’s arm during the Sol 103 plan.  The CI team had written down their intentions for their observations, using some of the targets we had selected during Sol 101 planning as guideposts for their discussions.  Everyone on shift agreed that it was a new and odd combination of multiple meetings proceeding simultaneously, but we understood that our ability to adapt to this new method of operating rovers remotely was a big part of the ROASTT itself. 

After discussing the Sol 102 outline, I signed off and finished making notes on my TSL report for the next shift of tactical team members to read.  In a few days I would participate as a Payload Downlink Lead (PDL) for the Mastcam-Z instrument, confirming that the planned images were acquired correctly and that the instrument was ready to continue acquiring more pictures.  But for now, Sol 101 was in the books, the first day of our test.  The field team in the desert would spend tomorrow taking the images we requested, and tactical planning for Sol 102 would start the day after.  After Sol 102, the team would repeat the process for the next two weeks.  And two Februarys from now on a craggy plain in Jezero Crater, we’ll get to do it for real.