The Mars Science Laboratory is more than the biggest rolling science laboratory ever put on another planet. It's a systems engineering—and product development—triumph.


click to enlarge
This graphic shows the locations of the 17 cameras on NASA's Curiosity rover. Installed at NASA's Jet Propulsion Laboratory, these cameras are crucial to a large number the rover's navigation and science missions. Engineers also designed the robotic arm to use these cameras to position itself when drilling and taking samples. Image: NASA/JPL-Caltech   

The world had seen a rover land on Mars before. In 1997, the Sojourner had done it. In 2004, the Mars Exploration Rovers did it. But the Curiosity, which recently analyzed its first drilled rock sample on Mars, was a different animal. This rover was much bigger than previous rovers, and was far heavier and more complex than any wheeled vehicle NASA has put on another planet.

The public realized this, and was also fascinated by the dramatic landing. For most, the highlight of the Aug. 6, 2012, touchdown was the audacious entry, descent, and landing (EDL) procedure, a parachute-skycrane-airbag plan that seemingly offered a dozen different ways to fail. The successful landing was a relief to many, and the American public was treated to the sight of a mission control in full jubilation.

All eyes were soon on the rover. At NASA, the concern was how much the stress of travel had affected the careful engineering that many thousands of people had spent years refining just for this mission. How well would Curiosity function when its airbag deflated and the wheels took their first turns in the Martian soil?

As of December 2012, their worries have turned into work as the rover has begun its primary mission of studying the Martian surface in order to learn more about how the planet formed and what sort of activity—up to and including the possibility of ancient life—occurred on Mars.

Conducting this sort of science requires a highly capable rover, and more than any rover or lander in NASA’s history, Curiosity can function at a high level as a remote laboratory, one that is able to conduct high-level optical, thermal, and chemical analyses on a variety of geologic samples. For the effort and skill NASA, its collaborators, and thousands of engineers and scientists around the world have devoted to the successful development of Curiosity, the editors of R&D Magazine present its highest award for innovation, the 2012 Innovator of the Year Award, to the NASA Mars Science Laboratory team and its partners.

The problem of scale
Mars is one of the most compelling places in the solar system for geology because of its relative accessibility and the overwhelming evidence that some interesting geological activity has occurred in its past. The previous rovers had been highly successful according to their designs, but they only whetted the appetite of planetary scientists, who hadn’t had a chance to examine Martian soil with sensitive scientific instrument, such as X-ray spectrometers.

“Scientists, of course, will ask for as much capability as they can. I think that was another one of the challenges we faced with this mission, because we were so successful with the Mars Exploration Rovers, we were eager to do our next big rover and eager to better ourselves,” says Rick Welch, a system engineer at NASA’s Jet Propulsion Laboratory (JPL), Pasadena, Calif., and one of MSL’s tactical mission managers. Welch has been at JPL for 20 years, was involved with the Sojourner mission in 1997, and has since been involved on all Martian rover missions.

“I was still engaged in Mars Exploration Rovers while the Mars Science Laboratory was really in its formulation stage,” says Welch. “At the time its first design, in 2003, MSL was a big rover that was designed to carry large payloads. The payloads really drove the size of the rover, and the initial designs showed that it was very big and capable, with two robotic arms and two power sources.”

Cost and weight factors soon shaved away the extra arm and power source, but the MSL rover still presented some significant challenges, particularly with regard to size. Sojourner was 25 pounds and the Mars Exploration Rovers were 350 pounds each. The Curiosity is 2,000 pounds.

“With that change in scale there are a lot of things that came to light. The total amount of energy that had to travel through the rover meant that total cable length was well over 20 m from one end of the rover to the end of the robotic arm,” says Welch. Large actuators that operated the arm require a lot of electrical power, causing major power drops along its length.

The other factor that dictated a larger rover was the need to get samples out of hard rock material. The best way to do this, they felt, was to use much the same tools that geologists on Earth use: a hand lens, a drill, and a variety of sample analysis instruments, such as a spectrometer.

This raised the difficulty levels, and the risk, for the team. In addition to functioning properly on Mars, with its low temperatures and low atmospheric pressures, the instruments would have to endure the stress of launching from Earth and landing on Mars. High radiation would be a factor, too, when it came to designing adequate protection in the spacecraft. Scientific instruments would have to be radiation-hardened for everyday use on Mars, as well.

“That’s why a lot of our electronics are so customized and relatively out-of-date compared to everyday computers, because we’re using radiation-hardened components,” says Welch.

The weight was not initially a concern when flight engineers were developing their complex EDL landing approach. But the increase in rover mass caused airbag failures during testing on Earth, which made engineers realize there were limits with that approach.

The relatively thin atmosphere and moderate gravity of Mars presented a major headache for flight engineers. Curiosity was too heavy to simply drop to the surface on airbags. The atmosphere isn’t dense enough to allow a parachute to carry it down. The hybrid approach that involved a first-stage parachute, a second-stage tether seconds before landing, and an airbag cushion third stage, was a necessary complication.

“They came up with a system that engineers projected would be 95% reliable on the surface, but, still, the landing is crazy to watch,” says Welch.

The importance of systems engineering
Curiosity runs on the fissile decay of plutonium, and can travel up to 90 m per hour over obstacles up to 75 cm high. Curiosity’s average speed so far in its mission has been much slower; its scientific throughout for sample analysis is low. However, conducting science remotely at a distance of millions of miles is by necessity highly deliberate. According to JPL’s Dan Limonadi, flight systems engineer, the process is a “delicate dance” as the rover’s operators operate one subsystem—the wheels, the arm, the camera—one at time to ensure everything is operating smoothly and properly.

Limonadi has been on the mission since 2004 as the lead systems engineer for surface sampling. His responsibility leading up to the 2011 launch was to ensure that the CheMin instrument was properly integrated with the Mars Science Laboratory rover and mission ready for sampling on the Martian surface. Now, he’s responsible for a wide variety of day-to-day rover operations.

“I’m juggling three balls with this project,” says Limonadi. In addition to managing the integration of systems, such as the Chemistry and Mineralogy instrument, he assists with the hardware subsystem team and the instrument team, ensuring hardware functions properly. He is also in charge of systems integration and watches to see that all systems work properly together.

“The main challenge is that much of this hardware has never really done this type of activity before,” says Limonadi. “Actually drilling a hole on another planet is something that’s rarely been done.”

A characterization phase, in which all systems are tested in small increments to reveal any glitches or potential problems, lasted for many weeks after touchdown. Only in the past few weeks have the first drilling samples been made.

Limonadi joined the MSL mission after working on the design of the MER rovers. His previous experience was used in his role as tailored accommodation engineer to figure out how to integrate the large suite of instrument on board Curiosity. Many of the instruments installed on Curiosity were developed elsewhere. The large and complex Science at Mars instrument, for example—which is actually three scientific instruments in one—was built at Goddard Space Flight Center. Others were designed or built in Europe or Russia.

According to Welch, this type of collaboration is common for JPL, and is a benefit because it brings design viewpoints from around the world. But at NASA, overall development decisions needed to be made at a relatively early stage to ensure mission parameters were met.

Technicians and engineers inside a clean room at NASA's Jet Propulsion Laboratory (JPL), Pasadena, Calif., prepare to install the Sample Analysis at Mars (SAM) instrument packaged into the mission's Mars rover, Curiosity, a few months before launch in 2011. Designed by NASA's Goddard Space Flight Center, SAM is the largest of Curiosity's 10 science instruments. Image Credit: NASA/JPL-Caltech“At the time, people didn’t really know whether the instruments would do the sample analysis processes themselves,” says Welch. “One of the early design studies asked ‘Is it better to provide a facility for all of the instruments as opposed to let individuals try to design things on their own?’”

That’s a common theme for a lot of NASA missions, as planners try to decide whether to provide a common resource so the instrument makers are not trying to duplicate efforts. As a result, the MSL project was designed as way to prevent conflicts of function or effort; and JPL was tasked with the job of integrating the instruments package, providing a stable home for scientific instruments.

“JPL has always had a very strong systems engineering effort because of the interactions between different systems, particularly for interplanetary mission like Mars, but we’ll have the same challenges if we go to Europa or other places,” says Welch.

Limonadi’s involvement with the installation of CheMin is indicative of the systems engineering challenges NASA engineers faced with the MSL project. The CheMin instrument is a platform for extraterrestrial X-ray diffraction analysis and has its roots in the work of NASA’s Ames Laboratory’s, Ames, Iowa, David Blake, who began working two decades ago on a compact, portable X-ray diffraction instrument (XRD) suitable for use in space. He and NASA fellow Philippe Sarrazin developed a powder vibration system that allowed the XRD device to be far more sensitive and accurate when used with samples of varying grain sizes. This held importance for the mission because Curiosity’s drilling mechanism would not be able to generate consistently sized powder. This technology, which won an R&D 100 Award in 1999 from R&D Magazine, was licensed to InXitu Inc., Campbell, Calif.

Even so, the terrestrial XRD was far from ready for duty in space, says Limonadi. CheMin required a number of improvements. Despite the frigidity of Martian air, the proximity of other instruments meant that the CheMin’s charge-coupled device (CCD) that collects the X-ray information would have trouble maintaining a temperature cool enough to operate efficiently. Limonadi and other engineers helped adapt a Sterling engine cryo-cooler made by Ricor originally invented for use on heat-seeking missiles.

“Another challenge we had was getting a X-ray source supplier,” says Limonadi. JPL couldn’t find anybody who could meet the required specifications, so much of the X-ray source work was done at JPL.

“Getting space-qualified versions of what you can find on the shelf on Earth is difficult, and often have to be re-engineered. Doing this work in time for launch was a real challenge,” says Limonadi.

Curiosity’s heavy arm
One advantage the MSL team did have is experience. In addition to the Viking landings in the 1970s, many of the same NASA engineers that designed the Mars Exploration Rovers were also taking part in the MSL mission. As a result, they understand the stresses that could damage instrumentation, materials, and electronics.

“As least from an accommodations perspective, we had been to Mars before, we knew the environment, and so we were able to design the rover to provide a nice thermal environment,” says Limonadi. The rover is equipped with radioisotopic heat exchangers that deliver 1,800 W of heat and 110 W of power. The avionics equipment is bolted to a fluid loop control interface that guarantees a warm environment for those systems that require it. However, many of the on board instruments, such as the CheMin and the SAM, require operating temperatures far colder.

“Electronics function well at 0 to 20 C, but a lot of instruments want to be super-cold, and at the business end they often like -60 C and colder. This raised a lot of standard engineering questions about how to balance that heat over a large area,” says Limonadi.

Solving thermal issues required creative thinking and careful planning, and the rover’s robotic arm required innovative packaging.

“The mission criteria was an arm that could pick itself up and drill into solid rock with 100 pounds of force, stay stable while drilling, and be accurate enough to place relatively precisely,” says Limonadi. It also had to possess enough degrees of freedom to deliver 40 to 50 mg of sample material to the holders for several different instruments.

Adding complexity was the fact that a number of systems had to be located at the business end of the arm: a mechanical device for scooping and collecting drilled samples, a dust removal tool, the Alpha X-ray Portable Spectrometer (APXS), and the Mars Hand Lens Imager (MAHLI). The resulting design is 2 m long and weighs 100 kg, with much of weight at the end of the arm.

The final design of the titanium arm features two joints at the shoulder, one at the elbow, and two at the wrist. Each joint moves with a cold-tolerant actuator custom-built for the mission. When testing the arm on Earth, Limonadi recalls, the arm could barely hold its own weight. On Mars, the arm would operate to specification.

The design of the guidance system highlights the compromises that NASA’s engineers had to make to hold down both expense and complexity. On Earth, targeting can be as simple as using a device equipped with a GPS chip and software. On Mars, the rover finds its way around the surface with complex software.

“This is different from mobility software. For the arm, we just didn’t have the computing power and time to develop special path-planning software,” says Limonadi. The arm guides itself through a system of stereo cameras attached to the rover mast. The cameras work together to close a loop on a stereo model of the world. This helps the arm designate a target, which is fed into high-level software that guides the arm into place. Gravity readings help them align a vector on the turret to begin drilling.

Complexity refines strategies
At one time, nearly half of the staff at JPL was involved in the Mars Science Laboratory development. Significant teams at nearly all NASA sites, including Ames Laboratory, Glenn Research Center, and Goddard Space Flight Laboratory, were involved in the mission. Thousands more from outside companies, such as Lockheed Martin, were involved in the logistical spaceflight engineering, and a large number of outside engineering efforts to develop scientific instrumentation and communication systems took place outside the space agency walls.

Welch agrees that JPL and other NASA sites, when working with many partners, have to strike a careful balance.

“We don’t have infinite resources. On the sampling system, the sample processing module was originally mounted back on the rover itself. We looked at saving costs and complexity. We had to make a tough decision about coupling the drill systems in the arm,” he says. This effort was further complicated when the original 2009 launch date for MSL was pushed back to 2011. If Limonadi had known a 2011 launch date would happen, he believes he could have done more development, particularly on the drill design, which had been finalized for the 2009 date.

“The classic problem that NASA has is it must try to develop something that has never or rarely been done before. That introduces a lot of cost pressure,” Limonadi says. “Then you have to make it happen with the people you have, and the team must understand the priorities.”