How Lack of Information Sharing Jeopardized the NASA/ESA Cassini/Huygens Mission to Saturn
Lead Author: Brian White
This article describes a deep space mission where more forthright information exchanges between teamed but rival agencies could have preserved the original plan as well as saved much time and money. The topic may be of particular interest to those involved in institutional collaborations where there are vested interests in protecting rather than sharing information.
For addition information, refer to the closely related topics of Information Management, Organizing Business and Enterprises to Perform Systems Engineering and Fundamentals of Services.
Before the “Faster, Better, Cheaper” philosophy introduced in the 1990s, the United States National Aeronautics and Space Administration (NASA) focused on three classes of unmanned space missions. In order of increasing cost, these were the Discovery, New Frontiers, and Flagship programs. Flagship programs typically cost more than $1B, and included the Voyager (outer planets), Galileo (Jupiter), Cassini-Huygens (Saturn), Mars Science Laboratory (Mars), and the James Webb Space Telescope. (Wall 2012)
The concept of the Cassini-Huygens mission was initiated in 1982 as the result of a working group formed by the National Academy of Sciences and the European Science Foundation. This group sought opportunities for joint space missions; several subsequent reports endorsed the working group’s concept of a Saturn orbiter coupled with a Titan (Saturn’s largest moon) lander. (Russell 2003, p. 61)
By 1988, NASA was politically motivated to reverse earlier tensions with the European Space Agency (ESA) by engaging in a joint mission. Cassini-Huygens was seen as a mechanism to achieve this goal, and the cooperation between NASA and ESA helped the program survive potential budget cuts (since the U.S. was obligated to match ESA commitments). (Russell 2003, p. 62)
NASA and ESA approved the Cassini-Huygens program, and it proceeded under a traditional management approach. NASA built the Cassini orbiter (the largest and most complex unmanned space probe ever built) and the ESA constructed the Huygens lander. This partition of responsibility almost led to the failure of the Titan survey portion of the mission. Cassini (which would conduct a variety of scientific surveys of the Saturn planetary system) was expected to relay transmissions from Huygens to NASA’s Deep Space Network (DSN); however, the interface between the lander and orbiter was not well-managed and erroneous assumptions about how the orbiter/lander system would behave after separation nearly doomed the Titan exploration portion of the mission. (Oberg 2004)
The intent of the Titan survey portion of the Cassini-Huygens mission was that the Huygens lander would separate from the Cassini orbiter and commence a one-way, 2.5 hour descent into Titan’s atmosphere. Its modest transmitter would send data back to the orbiter, which would relay the information to Earth. (Oberg 2004, p. 30) This effectively made the radio link between the two spacecraft a single point of failure (SPOF) and one that was not well characterized.
Alenia Spazio SpA, the Italian communications vendor that built the radio system, overlooked the Doppler shift (approximately 38 kHz) (Oberg, 2004, p. 31) that would occur when Huygens separated from Cassini and began its descent (Oberg 2004, p. 38). The communications protocol was binary phase-key shifting: “[the] transmission system represents 1s and 0s by varying the phase of the outgoing carrier wave. Recovering these bits requires precise timing: in simple terms, Cassini’s receiver is designed to break the incoming signal into 8192 chunks every second. It determines the phase of each chunk compared with an unmodulated wave and outputs a 0 or a 1 accordingly”. (Oberg 2004, p. 31) The receiver was appropriately configured to compensate for the Doppler shift of the carrier wave but would be unable to adjust for the Doppler shift of the encoded data. “In effect, the shift would push the signal out of synch with the timing scheme used to recover data from the phase-modulated carrier.” (Oberg 2004, p. 33) Therefore, the communications system would be unable to decode the data from the lander and would then relay scrambled information to NASA. Because of the failure mechanism involved, the data would be completely unrecoverable.
Both Cassini and Huygens had been tested before launch; however, none of the testing accurately reflected the Doppler shift that would be experienced at this critical phase of the mission. An opportunity to conduct a full-scale, high-fidelity radio test was ignored due to budget constraints; the testing would have required disassembly and subsequent recertification of the probes. (Oberg, 2004, p. 30) Correcting this latent issue would have been trivial before the spacecraft were launched (via a minor firmware upgrade); (Oberg 2004, p. 33) once they were on the way to Saturn any corrective action would be severely limited and expensive.
Once the mission was underway, the probe coasted along its seven-year trajectory to Saturn and its moons. Claudio Sollazzo, the ESA ground operations manager, was uncomfortable with the untested communications system. He tasked Boris Smeds, an engineer with radio and telemetry experience, with finding a way to test the communications system using an Earth-generated signal. (Oberg 2004, p. 30)
Smeds spent six months developing the test protocols that would use Jet Propulsion Laboratory (JPL) ground stations and an exact duplicate of Huygens. Simulated telemetry would be broadcast from Earth to Cassini and relayed back; the test signal would vary in power level and Doppler shift to fully exercise the communications link and accurately reflect the anticipated parameters during Huygens’s descent into Titan’s atmosphere. (ESA 2005)
Smeds faced opposition to his test plans from those who felt it was unnecessary, but ultimately prevailed due to support from Sollazzo and Jean-Pierre Lebreton, the Huygens project scientist. More than two years after the mission was launched, Smeds traveled to a DSN site in California to conduct the test. (Oberg 2004, p. 31)
A test signal was broadcast, received by Cassini, re-transmitted to the DSN site, and relayed to ESA's European Space Operation Centre (ESOC) in Darmstadt, Germany for analysis. Testing had to be conducted when the orbiter was in the correct relative position in the sky; it was more than a quarter of a million miles away with a signal round-trip time of nearly an hour. The test immediately exposed an issue; the data stream was intermittently corrupted, with failures not correlated to the power level of the test signal. The first of two days of testing concluded with no clear root cause identified. (Oberg 2004, p. 31)
Even though the probe was far from its ultimate destination, many science teams were competing for time to communicate with it using the limited bandwidth available. The communications team would not be able to conduct another set of trials for several months. Smeds diagnosed the root cause of the problem; he felt it was the Doppler shifts induced in the simulated signal. However, the test plan did not include unshifted telemetry (an ironic oversight). He modified his test plan overnight and shortened the planned tests by 60%; this recovered sufficient time for him to inject an unshifted signal into the test protocols. (Oberg 2004, p. 32)
This unshifted signal did not suffer from the same degradation; however, other engineers resisted the diagnosis of the problem. Follow-up testing using probe mockups and other equipment ultimately convinced the ESA of the issue; this took an additional seven months. (Oberg 2004, p. 33)
By late 2000, ESA informed NASA of the latent failure of the communications link between Cassini and Huygens. Inquiry boards confirmed that Alenia Spazio had reused timing features of a communications system used on Earth-orbiting satellites (which did not have to compensate for Doppler shifts of this magnitude). (Oberg, 2004, p. 33) In addition, because NASA was considered a competitor, full specifications for the communications modules were not shared with JPL. The implementation of the communications protocols was in the system’s firmware; trivial to correct before launch, impossible to correct after. (ESA 2005)
A 40-man Huygens Recovery Task Force (HRTF) was created in early 2001 to investigate potential mitigation actions. Analysis showed that no amount of modification to the signal would prevent degradation; the team ultimately proposed changing the trajectory of Cassini to reduce the Doppler shift. (ESA 2005) Multiple studies were conducted to verify the efficacy of this remedy, and it ultimately allowed the mission to successfully complete the Titan survey.
Systems Engineering Practices
Space missions are particularly challenging; once the spacecraft is en route to its destination, it is completely isolated. No additional resources can be provided and repair (particularly for unmanned mission) can be impossible. Apollo 13’s crew barely survived the notable mishap on its mission because of the resources of the docked Lunar Excursion Module (LEM) and the resourcefulness of the ground control team’s experts. A less well-known failure occurred during the Galileo mission to Jupiter. After the Challenger disaster, NASA adopted safety standards that restricted the size of boosters carried in the Space Shuttle. (Renzetti 1995) Galileo was delayed while the Shuttles were grounded and Galileo’s trajectory was re-planned to include a Venus fly-by to accelerate and compensate for a smaller booster. Galileo’s main antenna failed to deploy; lubricant had evaporated during the extended unplanned storage (Evans 2003) and limited computer space led to the deletion of the antenna motor-reversing software to make room for thermal protection routines. When the antenna partially deployed, it was stuck in place with no way to re-furl and redeploy it. Engineers ultimately used an onboard tape recorder, revised transmission protocols, the available low-gain antenna, and ground-based upgrades to the DSN to save the mission. (Taylor, Cheung, and Seo 2002)
The Titan survey was ultimately successful because simulation techniques were able to verify the planned trajectory modifications and sufficient reaction mass was available to complete the necessary maneuvers. In addition, Smeds’s analysis gave the mission team the time it needed to fully diagnose the problem and develop and implement the remedy. If this test were conducted the day before the survey it would merely have given NASA and ESA advance warning of a disaster. The time provided enabled the mission planners to craft a trajectory that resolved the communication issue and then blended back into the original mission profile to preserve the balance of the Saturn fly-bys planned for Cassini. (Oberg 2004, p. 33)
The near-failure of the Cassini-Huygens survey of Titan was averted because a handful of dedicated systems engineers fought for and conducted relevant testing, exposed a latent defect, and did so early enough in the mission to allow for a recovery plan to be developed and executed. Root causes of the issue included politically-driven partitioning, poor interface management, overlooked contextual information, and a lack of appreciation for single-points-of-failure (SPOFs).
The desire to use a joint space mission as a mechanism for bringing NASA and ESA closer together (with the associated positive impact in foreign relations) introduced an unnecessary interface into the system. Interfaces must always be managed carefully; interfaces between organizations (particularly those that cross organizational or political borders) require extra effort and attention. Boeing and Airbus experienced similar issues during the development of the Boeing 787 and A380; international interfaces in the design activities and supply chains led to issues:
…every interface in nature has a surface energy. Creating a new surface (e.g., by cutting a block of steel into two pieces) consumes energy that is then bound up in that surface (or interface). Interfaces in human systems (or organizations), a critical aspect of complex systems such as these, also have costs in the effort to create and maintain them. Second, friction reduces performance. Carl von Clausewitz, the noted military strategist, defined friction as the disparity between the ideal performance of units, organizations, or systems, and their actual performance in real-world scenarios. One of the primary causes of friction is ambiguous or unclear information. Partitioning any system introduces friction at the interface. (Vinarcik 2014, p. 697)
Alenia Spazio SpA’s unclear understanding of the Doppler shift introduced by the planned relative trajectories of Huygens and Cassini during the Titan survey led it to reuse a component from Earth-orbiting satellites. Because it considered NASA a competitor and cloaked details of the communications system behind a veil of propriety, it prevented detection of this flaw in the design phase. (Oberg 2004, p. 33)
Because NASA and ESA did not identify this communication link as a critical SPOF, they both sacrificed pre-launch testing on the altar of expediency and cost-savings. This prevented detection and correction of the flaw before the mission was dispatched to Saturn. The resource cost of the later analysis and remedial action was non-trivial and if sufficient time and reaction mass had not been available the mission would have been compromised. It should be noted that a number of recent spacecraft failures are directly attributable to SPOFs (notably, the Mars Polar Lander (JPL 2000) and the Genesis sample return mission (GENESIS, 2005)). Effective SPOF detection and remediation must be a priority for any product development effort. More generally, early in the development process, significant emphasis should be placed on analyses focused on what might go wrong (“rainy day scenarios”) in addition to what is expected to go right (“sunny day scenarios”).
The success of the Huygens survey of Titan was built upon the foundation established by Boris Smeds by identifying the root cause of the design flaws in a critical communications link. This case study underscores the need for clear contextual understanding, robust interface management, representative testing, and proper characterization and management of SPOFs.
Evans, B. 2003. "The Galileo Trials." Spaceflight Now. Available: http://www.spaceflightnow.com/galileo/030921galileohistory.html.
GENESIS. 2005. "GENESIS Mishap Investigation Board Report Volume I." Washington, DC, USA: National Aeronautics and Space Administration (NASA).
JPL. 2000. "Report on the Loss of the Mars Polar Lander and Deep Space 2 Missions." Special Review Board. Pasadena, CA, USA: NASA Jet Propulsion Laboratory (JPL).
ESA. 2005. "Modest Hero Sparks Team Response." European Space Agency. Available: http://www.esa.int/Our_Activities/Operations/Modest_hero_sparks_team_response.
Oberg, J. 2004. "Titan Calling: How a Swedish engineer saved a once-in-a-lifetime mission to Saturn's mysterious moon." IEEE Spectrum. 1 October 2004, pp. 28-33.
Renzetti, D.N. 1995. "Advanced Systems Program and the Galileo Mission to Jupiter." The Evolution of Technology in the Deep Space Network: A History of the Advanced Systems Program. Available: http://deepspace.jpl.nasa.gov/technology/95_20/gll_case_study.html.
Russell, C. 2003. The Cassini-Huygens Mission: Volume 1: Overview, Objectives and Huygens Instrumentarium. Norwell, MA, USA: Kluwer Academic Publishers.
Taylor, J., K.-M. Cheung, and D. Seo. 2002. "Galileo Telecommunications. Article 5." DESCANSO Design and Performance Summary Series. Pasadena, CA, USA: NASA/Jet Propulsion Laboratory.
Vinarcik, M.J. 2014. "Airbus A380 and Boeing 787 — Contrast of Competing Architectures for Air Transportation," in Case Studies in System of Systems, Enterprise Systems, and Complex Systems Engineering, edited by A. Gorod et al. Boca Raton, FL, USA: CRC Press. p. 687-701.
Wall, M. 2012. "NASA Shelves Ambitious Flagship Missions to Other Planets." Space News. Available: http://www.space.com/14576-nasa-planetary-science-flagship-missions.html.