Acta Astronautica 57 (2005) 116 – 120

www.elsevier.com/locate/actaastro

The Mars exploration rover projectRichard A. Cook

Mars Exploration Rover Project, Jet Propulsion Laboratory, California, Institute of Technology, CA, USA

Abstract

The twin rovers of the Mars exploration rover project successfully landed on the surface of Mars in January 2004, andinitiated a highly successful science and exploration campaign. This marked the culmination of an unprecedented 4-yeareffort to design, build, launch, and operate two of the most complex planetary spacecraft ever built. The project was startedin the aftermath of the 1999 Mars mission failures, and was commissioned to take advantage of the highly advantageous2003 opportunity. The development schedule from project start to launch was only 35 months, so schedule management wasthe most significant challenge facing the project. This problem was compounded when early assumptions about the extent ofdesign heritage from Mars Pathfinder proved to be flawed. The project derived a number of useful lessons learned in solvingthese challenges that can be applied to future missions.© 2005 Published by Elsevier Ltd.

1. Project formulation

The Mars exploration rover (MER) project wasstarted in May 2000 during the turbulent period fol-lowing the loss of the Mars climate orbiter and Marspolar lander missions. NASA initiated an effort torestructure the Mars exploration program, but recog-nized the need to take rapid advantage of the excellent2003 opportunity. Two missions were identified asstrong candidates for 2003: a large orbiter to follow onthe Mars Odyssey mission or a rover mission based onthe successful Mars Pathfinder mission. After a shortstudy period, NASA decided to proceed ahead with

0094-5765/$ - see front matter © 2005 Published by Elsevier Ltd.doi:10.1016/j.actaastro.2005.03.008

the rover mission in July 2000. The rationale for thisdecision included some of the following elements:

• The 2003 opportunity was ideal for a lander mis-sion, with good geometry and energetic consid-erations for both launch and arrival.

• The orbiter concept would be very difficult pro-grammatically because neither the spacecraftprovider nor payload elements had been selected.

• The rover concept was relatively mature becauseof the Mars Pathfinder heritage and becausethe science payload could be inherited from theAthena rover study.

R.A. Cook / Acta Astronautica 57 (2005) 116–120 117

The recent failure of Mars Polar Lander, however,had made a number of people concerned about thewisdom of sending a single vehicle into the harshand unpredictable Martian environment. This concernresulted in a redirection of the project to build andlaunch two identical rovers to reduce risk and improvescience return.

Thus, the MER project was initiated with the ob-jective of placing two highly capable rovers, on thesurface of Mars with primary purpose to look forgeologic evidence of ancient water. Each rover wouldcarry a payload package consisting of a stereo pair ofhigh-resolution color cameras (PanCam), a miniaturethermal emission spectrometer (MiniTES) similar tothe instrument flown on the Mars Global Surveyorspacecraft, a Mossbauer spectrometer contributed bythe Max Planck Institute for Chemistry in Germany,an alpha particle X-ray spectrometer similar to theinstrument flown on the Mars Pathfinder Sojournerrover (also contributed by the Max Planck Institute),a microscopic imager, a set of magnets for magneticproperties assessment, and a Rock Abrasion Tool toenable sub-surface rock analysis. The key level 1requirements negotiated with the Mars explorationprogram were:

• Launch two rovers to Mars in 2003 opportunityon Delta ll-class launch vehicle.

• Provide real-time communications during Marsentry and landing to support fault reconstruction.

• Land and operate solar powered rovers in latitudeband±10◦ around the subsolar point.

• Operate each rover for at least 90 sols.• Utilize both X-band direct to Earth and UHF

communications through Odyssey.• Drive the rovers to at least eight separate loca-

tions and investigate geologic context and diver-sity.

• Drive more than 600 m on at least one rover.

2. Implementation challenges

The planned and actual implementation chronolo-gies of the MER project are shown inFig. 1. Duringthe early implementation phase, the project team fo-cused on validating the basic mission concept. Thiseffort led to the realization that the concept and under-

lying programmatic assumptions were not consistent.MER was based on the presumptions that the MarsPathfinder entry, descent, and landing approach couldbe inherited and that a large and capable rover wouldfit within the technical resource (mass, power, vol-ume) envelope provided by the Pathfinder landing sys-tem. Both assumptions proved to be wrong. The MarsPathfinder EDL system was developed in a faster, bet-ter, cheaper era and while the resulting mission wassuccessful, there were aspects of the development pro-cess that were not appropriate for a high visibility,post-failure effort. In addition, the environmental as-sumptions around which the Pathfinder system wasdesigned (winds, site altitude, Earth-Mars geometry)were not valid for the 2003 opportunity. These changesalone would have forced some changes to the build-to-print approach.

The more significant impact on heritage, however,came from the mismatch between the Pathfinderdelivery systems resource capability and the require-ments of the Athena rover. An assumption madeduring concept development was that the projectwould operate in a capabilities driven/resource con-strained environment. In that model, the rover, sci-ence payload, and mission return would be descopedto fit within the heritage envelope. The paradigmestablished by the Mars program however, was ascience driven model. Thus, the project had to re-design the EDL system to rectify the capabilitymismatch and preserve the science return. The ba-sic architecture (parachute, airbags, RAD rockets)remained similar (which proved to be critically im-portant), but the build-to-print programmatic planwas broken.

The project recognized this fundamental program-matic flaw in late 2000 and developed a technical re-covery plan by early the following year. Significantchanges were made to the design of the parachute,terminal descent rocket system, lander structure, andairbags. The parachute and airbags were modified toincrease the landed mass capability, while the landerstructure was changed from an all aluminum design toa composite approach to reduce mass. Additional ter-minal descent rockets were added to control horizon-tal landing velocity and decrease dependence on land-ing site winds. All four efforts became significant de-velopment challenges that aggravated an already tightdevelopment schedule.

118 R.A. Cook / Acta Astronautica 57 (2005) 116–120

5/00 8/01 5/0411/00

PDR

CDRStartATLO

ShipTo KSC Launch

2/02 6/03

Landing EOMMCR

CϕBA D E

1/042/03

5/00 8/01 5/0411/00

PDR CDR

StartATLO(EM)

ShipTo KSC Launch

3/02 6/03

Landing EOMMCR

CBA D E

1/042/03

The Plan

The Reality

PDRFltDel

8/022/01

ϕ ϕ ϕ ϕ

ϕϕϕϕϕ

Fig. 1. MER implementation chronology.

The cost and schedule implications of the loss ofheritage were not fully realized until several monthslater, near the time of the project critical design review.The net result of the change in the technical baselinewas the loss of 4–5 months on the critical path. Theproject invested significant budget resources to buyback some of that lost schedule, and had to adopt atwo shift, 6 day a week plan for spacecraft assembly,test, and launch operations (ATLO). The change inthe ATLO plan, in particular, allowed us to make upvirtually all of the schedule slip.

3. Quality challenges

The combination of the tight schedule and the reac-tion to the Mars 98/99 failures meant that the projecthad an extreme focus on maintaining acceptable qual-ity and mission risk. The project’s philosophy was toassure a risk level commensurate with previous classA flagship missions (Cassini, Galileo, etc.). This didnot mean, however, that it was okay to take risks thathad been acceptable in the past. The project team eval-uated the technical design, programmatic plans, andverification approach from first principles to determinethe overall risk posture. A good example is the over-all entry, descent, and landing system. Although thereturned out to be limited component heritage from MarsPathfinder, the system architecture remained similar.

The project, however, did not take any short cuts ineither the design process or verification program. Theresult is that we went back and challenged many ofthe assumptions and conclusions reached by the MarsPathfinder project. In a few limited cases, we founddesign problems or oversights that were not relevantto the smaller Pathfinder design, but could have beenproblematic for MER.

The tight schedule also caused a great deal of qual-ity concern because the potential existed to shortcutthe verification program to make it to the launch pad.A clear lesson from Mars ’98 was that a strictly rigor-ous verification and validation program is a mandatorycomponent of mission success. This is even more thecase on a schedule constrained effort like MER, be-cause a robust test program is the last line of defensefor a rapid design and development effort. To achievethis robustness, the project dedicated significant re-sources (personnel, hardware, budget) to the test pro-gram. In addition to the two flight spacecraft, a totalof four flight-like testbeds were developed and usedfor V&V. A very comprehensive environmental testprogram was conducted involving nearly 30 separatesystem level tests (solar thermal vac, launch dynam-ics, landing dynamics, EMC, etc.). A large test teamwas assembled from the design and development staffto ensure adequate cradle-to-grave knowledge.

The content of the test program was managed us-ing a very detailed incompressible test list (ITL).

R.A. Cook / Acta Astronautica 57 (2005) 116–120 119

Standard JPL practice is to define at a high level theset of mandatory tests which need to be performed onthe flight spacecraft prior to launch. Because of theMER schedule pressure, the project and institutiondecided to greatly expand the level of detail in the ITLand formalize the process for managing the list. Thescope of the list was also increased to include verifi-cation efforts that did not occur on the flight vehicle.This was particularly important given the criticalityof the testbed effort in verifying flight software func-tionality. The ITL was generated before the start ofthe verification program and was formally agreed toby both JPL and laboratory management. A smallnumber of waivers to the list were approved duringthe execution of the test program, but the vast major-ity of the tests were executed as planned. For eachwaiver, independent technical analysis was performedto assess the risk prior to institutional approval.

4. Transparency

The obvious challenges and importance of theMER project to both NASA and JPL meant that theproject was always under a great deal of internaland external scrutiny. This was a marked differencefrom the faster–better–cheaper era, when adoption ofa “skunkworks” approach was standard. Managingthe impact of this greater level of oversight was asignificant challenge to the project leadership team.The project’s approach to this effort was to developand maintain an extremely open and transparent re-lationship with the science team and NASA/JPLmanagement. This transparency requires overhead interms of communications and interactions, and alsocreates a trusting environment. This trust was criticalduring the periods when the project’s technical andprogrammatic challenges appeared insurmountable.Fortunately, both the science team and higher levelmanagement developed a strong belief in the team’sability to resolve issues—and this confidence turnedout to be well founded.

5. Technical challenges

The project faced a number of technical challengesin addition to the previously discussed programmatic

issues. Many of the technical issues were related tomass and the early disconnect between EDL systemcapabilities and rover needs. The project developed anew EDL system, and managing the mass of that ef-fort turned out to be a significant effort. A weeklymass council was conducted through most of the de-velopment effort to assess the mass status, identifygrowth areas, exercise descope options and trade re-sources (budget, schedule, mass, volume). As previ-ously mentioned, the limited mass margin forced theproject to optimize the design of many components.Many structural elements were fabricated using com-posite materials, and the project faced a number ofdifficult challenges, designing and building complexcomposite structures with many difficult interfaces.

Another significant challenge faced by MER wasthe difficulty in getting the rover off the lander ontothe surface of Mars. This problem was not really ad-dressed by Mars Pathfinder because the lander con-tained most of the mission functionality (the Sojournerrover effort solved the egress problem at a much re-duced scale). The challenge of driving a 400 kg roverover a tilted lander platform 40–50 cm off the groundwas not identified as a major concern early on. Even-tually, however, it became apparent that this problemwould drive the rover’s mobility capability beyondwhat was required to drive on Mars. This problem isaggravated by the presence of the deflated airbags be-low the lander petals. To mitigate this issue, a set ofspecial egress ramps were added to the design to aidthe rover during the egress activity. Nevertheless, thechallenges seen during the first few days of the mis-sion (particularly on spirit) showed that the design ar-chitecture made this problem a major challenge.

6. Key lessons learned

The major lessons learned on the MER project aremostly reflected in the challenges described above.The early history of the project demonstrates how im-portant it is to have an adequate program formula-tion effort. The lack of alignment between the projectteam and program office on the project’s developmentparadigm (science driven vs. capability/cost driven)could have been rectified with a longer formulationperiod. The same is true of the disconnect betweenthe capabilities of the inherited EDL system and the

120 R.A. Cook / Acta Astronautica 57 (2005) 116–120

rover’s needs. A longer and more complete formu-lation effort would have resulted in a better under-standing of the baseline schedule and budget, andwould undoubtedly reduce both the budget overrunand schedule slip. Of course, it was not really possi-ble to implement this lesson on MER due to the shorttime between project start (May 2000) and launch(June 2003).

Lesson: Provide for an adequate program formula-tion period to ensure that appropriate alignment occursbetween project paradigm, science objectives, techni-cal margins, and heritage assumptions.

Although challenging in some respects, the effort tobuild and fly two vehicles was extremely valuable. Thevalue of a dual mission from a mission risk/return per-spective is obvious. The project was able to leveragethe investment in the design/build effort to both reducethe landing risk and increase the science return. Thetwo missions also decreases the science risk becauseit allowed us to go to two scientifically diverse sites.The value of a dual mission from a development riskis not as easily assessed, but our experience is that it isequally important. It was certainly difficult to develop,build and test both flight units, but the advantage interms of hardware richness was worth it. The com-bination of the two flight vehicles and four testbedsmeant that the project could perform a very strong testprogram. By spreading V&V efforts through all avail-able venues, we greatly exceeded what would havebeen possible on a single vehicle. This improved op-portunity for testing undoubtedly led to discoveringmore problems and improving the overall quality ofthe product. This outweighed the logistical difficultiesof getting all the hardware built.

Lesson: Adopt a hardware rich development effortthrough multiple flight units or testbeds. Manage thelogistical difficulties in producing this hardware, andmake maximum use of this extra hardware in a robusttest program.

A number of factors contributed to the success ofthe MER mission. Of these, none was more impor-tant than the quality of the team. JPL put in placea strong and capable team at the beginning of theproject, and brought in additional senior personnelwhen the project was faced with significant develop-ment challenges. The project adopted a cradle to grave

staffing approach to ensure that the key design anddevelopment personnel stayed through test and flightoperations. This approach meant that the project couldretain important corporate heritage and eliminate in-efficient (and risky) personnel transitions. The projectalso made a specific effort to obtain a good mix of ex-perienced senior personnel and energetic newcomers.

Lesson: Staff the project with a strong & diverseteam and keep them from cradle to grave.

MER not only extended the capabilities of the MarsPathfinder EDL architecture, but also demonstratedthat further expansion is somewhat problematic. Theairbags had to be significantly upgraded to handleboth the increased rover mass and the larger wind-induced horizontal landing velocity. Further payloadmass growth would be difficult to accommodate be-cause the stroke capability is limited by the basic ge-ometry of the airbag system. In addition, the volumeinside the tetrahedron lander was quite limited, andforced the rover design to include a complex set of de-ployments. Adding more mass inside the lander wouldbe very difficult because of the limited volume mar-gin. Finally, the airbag architecture makes the task ofrover egress much more complex. Future EDL sys-tems should make sure that the task of safe landingincludes getting the payload all of the way onto thesurface of Mars.

Lesson: Further extension of the Mars Pathfinder/MER EDL architecture is likely to be problematic. Anew generation landing system should be developedto improve scalability and provide a robust approachto rover egress.

7. Summary

Despite the overall success of the Mars explorationrover missions, there were a number of challengesthat made the project exceptionally difficult to imple-ment. These problems include issues with the basicmission concept, a highly aggressive schedule, and asignificantly modified entry, descent, and landing sys-tem. Fortunately, these problems were successfully re-solved and the mission was executed as planned. Nev-ertheless, a number of useful lessons were learned thatshould prove to be useful for both future Mars rovermissions and general space exploration efforts.