Special Delivery

The 40-member team at NASA’S Jet Propulsion Laboratory that developed the Sky Crane landing system for the Mars Science Laboratory Curiosity rover is this year’s winner of the National Air and Space Museum’s trophy for current achievement. The team began work on the entry, descent, and landing procedures and design in 2003; the rover touched down on Mars on August 6, 2012. Team leader Adam Steltzner spoke with Executive Editor Paul Hoversten in February.

Air & Space: How was the Sky Crane developed?

Steltzner: The Sky Crane came into existence over several years, in fits and starts, with several contributions from different people. Its final germ happened in a great brainstorming session in the fall of 2003. In 1999-2000, after the loss of Mars Polar Lander, there were some teams working to try and understand how we would land the rover for the Mars Sample Return, which then was on the books for 2003 or 2005.  It had been planned to use a legged lander, like the one that had failed on MPL, but the mishap investigation helped underscore some of the weaknesses of a legged lander system. So the teams were looking at [other] ways of delivering a large lander.

There were several ideas out there, one of which was called Rover on a Rope. If you were to imagine the Mars Pathfinder or Mars Exploration Rover landing system and just strip away the airbags and have the rover naked at the end of the bridle, it was akin to that. But that idea was discarded by the teams as being too unstable and unusable.

The method that was chosen was something called the Pallet Lander, in which you take the legged lander and you give it six legs instead of three and you spread them out very flat to make the thing very stable. So the MSL went ahead with Pallet Lander. Unfortunately, we were struggling with the Pallet Lander [because it was too unstable.] We couldn’t use airbags or legged landers, but the experiences from them were there for us to understand how to innovate. We were forced into innovating by the laws of physics but it was the experience of the past that allowed us to know how to make a system that was as successful as it was.

So in the fall of 2003, we got everybody together for a big brainstorming session and threw out on the table everything that we’d previously considered, whether we’d rejected it or not. And we tried to work our way through to get past this logjam we were in with the pallet. [The final design] came out of that brainstorming session and some modifications of the ideas from Rover on a Rope, but with some very important additions, like leaving the parachute behind when the two bodies [the rover and its rocket-powered descent stage] are still attached, and waiting until the last minute in vertical flight to do the rover deployment. Those ideas really were the germs that made the Sky Crane happen.

How many people were in that brainstorming session?

There were about a dozen in the room. Naturally, people always want to ask who invented the Sky Crane. And my short answer is, none of us and all of us.

What was it like during the Seven Minutes of Terror it took for the rover to reach the surface?

I had the better part of a decade of my life invested in something that would all go down in the span of seven minutes. The number of things that had to go right to see the fruits of my decade ripen is remarkable. Thousands of lines of code, hundreds of devices, almost all of them being mission-critical, so it is terrifying. There was an interesting numerology for us when we were landing, because Mars was far enough away from the Earth that it took about 14 minutes for the signal to get from Mars to Earth. So it’s not only seven minutes of terror from the top of the atmosphere down to the surface, but it’s also true that when we first see that first signal, the rover’s been alive or dead for seven minutes on the surface. So seven was rolling all around there. When you’re actually in the event, on landing night, everybody in the control room is just a spectator. The vehicle is flying itself and we’re along for the ride.

Which part of the landing sequence was most worrisome?

The thing that we felt, mathematically, was the single lowest reliability element was probably the parachute, and that’s just [because of] the intrinsic uncertainties associated with parachutes. We throw more than 10,000 troops out of airplanes each year, and largely it’s into a very controlled environment, and we have parachutes designed exactly for those conditions. But we still give them a second parachute because even with all those controls, the odds just aren’t good enough. That’s not the case when you’re moving supersonically in an uncertain atmosphere 10 kilometers above the surface of Mars. So when you do the numbers, you end up convincing yourself that the single highest device risk is the parachute.

But that’s not the thing I was worried about. I was worried we could have missed something on the Sky Crane. It was so new and different. It was the unknown unknowns I was most concerned about, some unappreciated feature of this new landing system that we did not absorb and was waiting to bite us. So on landing night, as the data clicked by, I became more and more anxious. I said ‘Oh my God, is it really going to happen just this easily?’ I was pretty wound up for those last 20 or 40 seconds.

And your reaction afterward?

Tremendous relief, tremendous exhilaration, and, frankly, a slight sense of surrealness. To work on something for the better part of a decade, and then to have it done. Regardless of the outcome. It’s awesome that it was done successfully, but, I mean, all of a sudden it’s over. It was such a build-up, so much of my life was invested, and then it’s now, well, it just happened, now we move on.

The Sky Crane will also be used to land a rover similar to Curiosity in 2020. Are you studying any improvements to the system?

We knew going in that we had some points where the design was not all that we would have hoped it could have been. We were left holding some concentrations of risk that we would have preferred not to have. For example, we were measuring only two components of our velocity in the Sky Crane maneuver with our radar. We wanted to measure three, but because of late antenna development challenges, we could not get the right view angles to do that. So we said, ‘Well, we understand the local gravity of Mars fairly well at the landing site, so we’ll just estimate the third component.’ As it turned out, we didn’t understand the gravity well enough. There was a gravity anomaly at Gale Crater, which is probably not surprising because there’s a huge crater there. And that meant that we had an error in our estimates and we landed much more slowly than we’d anticipated. If that error had been flipped around, we might have landed faster than anticipated and we might have hurt the rover.

That’s an example of something the team that is going to be flying 2020 is going to look long and hard at, whether to put on a third antenna. For the lay person, the Sky Crane will look identical, but there will be some subtle changes the team will make to strengthen its reliability.