Red and The Robots

The scraping of metal wheels on loose rocks and the clicking sounds of mechanical actuators alert me to the lunar rover’s presence before I see it. Turning, I come face to face with the robot as it emerges from a shallow ditch, its two mast-mounted camera “eyes” gazing at the ground, then tilting up to scout a way forward.

Less than five feet tall and three feet across, it’s an unassuming ’bot: a truncated pyramid plastered with solar panels, moving on four wheels tucked underneath. As it passes me, the rover steers off to the right and trundles slowly on a 500-yard trek toward its goal: a crude mockup of the Apollo 11 lunar lander base, spray-painted gold—an incongruous sight here on the banks of the Monongahela River in Pittsburgh.

In May 2010, a descendant of this rover is scheduled to visit the actual Apollo 11 landing site on the moon in an attempt to claim the $25 million Google Lunar X Prize for its creators, Astrobotic Technology. A spinoff from Carnegie Mellon University’s Field Robotics Center, Astrobotic is led by the center’s founder, William Whittaker, known to all as Red.

Every other Friday, this second generation prototype of Red Rover—as the robot is named—goes for a test run here at Carnegie’s Robot City, a 40-acre strip of shale and gravel at Pittsburgh’s last working steel mill, which closed in 1979. Coincidentally, that was the year Whittaker finished his Ph.D. at Carnegie Mellon University and began building autonomous machines that would eventually explore active volcanoes in Alaska, map coal mines in central Pennsylvania, and search for meteorites in Antarctica.

Now one of the world’s foremost roboticists, Whittaker, 60, recently added a racing title to his storied career. In November 2007 he won the $2 million top prize in the DARPA Urban Challenge with a robotic vehicle capable of driving itself in simulated traffic over a 60-mile course. His hard-charging Chevy Tahoe, Boss, outran all 11 finalists, winning the final event by 20 minutes.

I first met Whittaker during that race, which was held in the California desert town of Victorville. Draping a red CMU sweatshirt over his bald head to shade his fair skin from the Mojave sun, he peered out like some hooded wizard. A tall man, Whittaker habitually leans into conversations and speaks with a quiet intensity. That day, even as he was in hot pursuit of the DARPA prize, his mind was on the moon.

In fact, it has been for years. While NASA has focused on sending robots like Spirit and Opportunity to Mars, Whittaker says the moon has been “under-considered by the mainline space robotics community.” And although he has built several robots for NASA, none has made it off Earth, let alone to the moon. So when the X Prize Foundation and Google announced their lunar contest on September 13, 2007, within hours Whittaker dashed off a check for the first installment of the $10,000 registration fee and overnighted it to the foundation. “There was an immediate attraction” to the lunar competition, he says. “I saw it as being inspirational, visionary, a very bold step for robotics.” Within 24 hours he assembled “a dozen compatriots” for the project. Among them were David Wettergreen, a CMU associate professor of computing who has worked with Whittaker on exploration robots for 20 years, and Sam Harbaugh, a 1958 Carnegie grad and systems engineer, “basically retired,” who had returned to campus to help run all three DARPA challenges. Working with these and other experienced engineers are Whittaker’s students, ranging from freshmen to post-docs. Sixty have signed up for this semester’s robotics course to help design and build Red Rover and its lander.

On the business side, Whittaker got back in touch with David Gump, a space entrepreneur with whom he had worked in the 1990s on a commercial proposal, LunaCorp, to launch a rover to the Apollo 11 site. Within weeks of reviewing the Google prize requirements, he and Gump realized that making another try at the moon would require a new company, and capital. So Astrobotic was formed, with Gump as president.

For a team to claim the Google prize, its robot has to land on the lunar surface, travel at least 500 meters (about a third of a mile), and send high-definition images and data back to Earth within 24 hours. The first team to do so will win $20 million; bonus awards totaling $5 million are offered for extras such as photographing an artifact of previous lunar exploration, travelling more than 5,000 meters, and operating for a second (two-week) lunar day. To win the full award, the mission must be completed by the end of 2012, and 90 percent of the funding has to come from private sources. So far, 14 teams have announced their intention to compete for the prize.

Originally, Whittaker and crew targeted a landing at one of the moon’s poles; the reserves of water ice believed to exist there would be useful to future lunar explorers. But ultimately “cultural interest” drove the decision: Astrobotic now intends to touch down near the Apollo 11 landing site in the Sea of Tranquillity and head off on a “Tranquillity Trek”—visiting the site of the first moonwalks, an area about the size of a soccer field, and sending back photos and video in near-real time.

In order for the rover to photograph itself on the moon (another Google requirement), the camera team is positioning a large parabolic mirror on the robot’s side, much like a bus mirror. This should also yield a “money shot” showing sponsor logos, the rover, and (perhaps) Earth. (There’s also talk of having the rover’s bulldozer-like treads imprint a sponsor’s logo or other design in the lunar dust.)

Red Rover will roll up as close to the Apollo lander as possible without trampling any footprints, and with its zoom lens try to photograph the famous “We Came in Peace For All Mankind” plaque on the lander base.

Whittaker has no doubt that his rover will be up to the task. “The tough nuts are the precision landing and a soft landing,” he told the assembled Google Lunar X Prize teams and the press in May. “When we nail that, it’s an easy journey. No matter what it takes, the robot will get us there.”

WHITTAKER’S CONFIDENCE comes from a lifetime of working with machines. Born in 1948, he grew up mainly in Hollidaysburg, Pennsylvania, a railroad town nestled in a small valley near Altoona. His mother was a chemist who taught school, his father a World War II bombardier who later sold explosives for mining and road construction.

His parents encouraged him to roam, and by age six he was raiding the local junkyard for parts. One of his first constructions was a rocketship, with rudimentary propulsion cooked up from a chemistry set. At 16 he fixed up a Jaguar XK-120 (he ended up driving it for years), and he took a job swinging a hammer on the railroad lines, where he learned that bending iron requires as much finesse as brute force.

In the late 1960s, he left Pennsylvania for Princeton University, intending to study civil engineering, but interrupted his education to enlist in the Marines, one eye on the educational benefits. When Apollo 11 landed on the moon in July 1969, Whittaker was in basic training. He has no memory of the event; on Parris Island, South Carolina, he says, there was “no news in, no news out.”

After two years of service, he returned to Princeton, G.I. Bill in hand. Following graduation in 1973, he continued on to Carnegie Mellon, where in 1979 he received his doctorate. That year, the Three Mile Island nuclear reactor in Harrisburg had a partial meltdown. In response, Whittaker and his  colleagues began building  robots to monitor and clean the reactor’s contaminated basement. The experience spurred him to found CMU’s Field Robotics Center.

In the mid-1980s, space beckoned. When NASA initiated a new class of low-cost Discovery planetary missions, Whittaker began pitching proposals, but none succeeded. The space agency did, however, fund a meteorite-hunting robot, Zoë, which the CMU team operated in Antarctica and Chile’s Atacama desert. On another Antarctic expedition, a walking robot named Dante tried to rappel into an active volcano, but got stalled by a kink in its fiber optic cable. A later version, Dante II, descended into an Alaskan volcano, a simulation of the harsh conditions on other worlds.

Meanwhile, Whittaker continued building robots for dirty, dangerous, and difficult jobs on Earth. After the 2002 collapse of the Quecreek mine in central Pennsylvania, which trapped nine coal miners, Whittaker and his students built two subterranean robots, Groundhog and Ferret, to show that they could map mines and perhaps prevent future flooding accidents. In 2004, Whittaker entered the first DARPA challenge, which he won on his third try, in 2007. Not all of his machines have worked perfectly, but all have worked, and many have been built on a tight schedule.

On a visit to Robot City last summer, I saw the Astrobotic team putting the prototype Red Rover through rigorous testing. In the back seat of a van that serves as a makeshift mission control, CMU software engineer Nathaniel Fairfield and a colleague used three laptops to run the rover’s navigation, safety, and control systems. The operators, who also wrote the software, could see what the robot saw as it drove, with a five-second delay built in to simulate actual moon operations. In some situations, the rover will have to “think” for itself without human input—for example, when navigating a slope.

“When are we in trouble?” asks Fairfield, his eye on a horizon indicator that would show if the rover were tilting too much.

“At 20 degrees,” his copilot says.

“We’re okay then.” In Earth gravity, Fairfield explains, the rover stays balanced because of its weight. On the moon, it will be six times lighter, so a 20-degree tilt could place it in danger of tipping over.

The details for these and other operations are figured out in the three-story High Bay, the working heart of the Robotics Center, located back on the main CMU campus. There I see a small fleet of machines, all built by Whittaker and his colleagues. The meteorite-hunting Nomad stands sentry near the door, wearing the NASA “meatball” logo. A few yards away sits Zoë, the solar-powered rover that in 2004 roamed more than 120 miles through the Atacama desert. An automated boat, Sol, awaits a dip in nearby Schenley pond. It’s a robotic wonderland.

At the far end, around the corner, is Red Rover’s shop area. Nick Miller, who is pursuing his master’s degree, shows me around. Having run through its paces, Red Rover now sits atop a rack, its solar panels removed to reveal wires, a central processing unit, and motors. This second prototype will be replaced by a third before next March. The fourth iteration—the flight version—now exists only as drawings on easel boards around the shop. In a nearby storage room, lead mechanical engineer John Thornton, wearing cargo shorts, and his team are fabricating their own carbon fiber parts. Thornton, who also has a master’s from CMU, apprenticed at Boeing’s Phantom Works.

Clearly, Astrobotic’s operation relies heavily on student power. “Nothing great ever came from robotics in our corner of the world that hasn’t had the energetic core of youth right upfront,” Whittaker told a gathering of the X Prize teams last year. Still, he admits, when it comes to making decisions, “Red is the last word. I’m not being imperial or pushy. [But] is the program going to take another minute going down a blind alley?”

BACK AT ROBOT CITY, I sit in on a Technology Interchange Meeting, where the project partners are discussing today’s engineering issues. A visiting team from Lockheed Martin is getting its first close look at Astrobotic’s plan. Collectively, the Lockheed team members have decades of experience working on NASA planetary missions, from the Mariners of the 1960s to the current Mars Reconnaissance Orbiter. On the phone are representatives of Astrobotic’s other partners: Raytheon, ATK, and the University of Arizona scientists who are planning the details of the Tranquillity Trek.

One of the topics at hand is Red Rover’s mass. With all its gear and a lander package called Artemis, the rover is overweight—its mass is more than the intended launch vehicle (Astrobotic won’t disclose their rocket supplier of choice) can lift. Not a huge problem at this early stage, but something that will need to be addressed.

The discussion moves on to the precision landing, a key part of Astrobotic’s plan. The landing will rely on Raytheon missile guidance technology adapted from the company’s Exoatmospheric Kill Vehicle. As the rover-lander descends to the lunar surface, an onboard computer will update its flight path by comparing pre-loaded photos of the terrain to real-time pictures taken during the descent. The target ellipse is just 1,100 yards long and 330 yards wide—demanding a high degree of accuracy for a planetary lander.

Tom Gardner of Raytheon presents the numbers for the final descent. At around two and a half miles above the surface, a rocket will fire to slow the lander-rover from 7,900 feet per second to 330 feet per second. It’s the equivalent of slamming on the brakes and diving down to the surface. At 23 seconds before landing, the main engine or smaller side engines will burn again for 2.5 seconds, further reducing the speed and setting up the soft landing.

The Lockheed engineers are concerned about this phase of the mission. To them it sounds very risky, “like falling out of the sky,” as one puts it. Finally, Gardner says, “Look, guys, this is not a traditional spacecraft mission…. This is more like a missile operation.” At the end of the two-day visit, the lead Lockheed engineer remains skeptical, but has changed his tune a bit. His verdict: The mission plan is “right on the verge of being possible.” But, he cautions, “you have a whole series of nasty discoveries in front of you.” For Whittaker, that means there’s still plenty of work to do, but he considers the landing problem tractable. And once the rover arrives on the moon and rolls off on its journey, he says, “it’s nothing that hasn’t been done before” by an earlier CMU robot.

Except that now it has to be done on the moon, nearly a quarter-million miles away.

At the Jet Propulsion Laboratory in Pasadena, California, which has built all of NASA’s planetary rovers, Rob Manning has been watching the Google Lunar X Prize contenders with a mix of enthusiasm and skepticism. As the chief engineer for JPL’s Mars program, Manning has a keen sense of what it takes to land and rove around on another planet. “Building robots that fly off Earth, land on another body, then interactively explore in a highly hostile environment requires a dazzling array of skills and technologies,” he points out in an e-mail. “I hope that [Astrobotic] fully appreciates this reality. I live it every day.”

Manning holds the individual members of Whittaker’s team in high regard. “The good news is that they’ve got incredible talent,” he says. But, he adds, “this is outside the box. It’s not a car, it’s not the DARPA challenge, not a missile. It’s an all new thing—taking the best ideas from very different places and putting them together in a very weird, highly coupled way that’s got to work the first time.”

If Astrobotic can figure out how to test its systems in an integrated way on Earth, Manning thinks the missile-like landing concept can work, even with the ridiculously low (by JPL standards) $100 million budget. “I think they have a shot at it,” he says. When I tell him that Astrobotic hopes to raise enough money for two shots, doubling its chances of success, he is happy to hear it: “That’s wonderful…. I want them to win.”

Whittaker and his financial team are not counting on angels to bankroll them out of kindness. Their business plan is based on the proposition that, before NASA sets up a base on the moon, robots will be making maps and collecting data. If the Astrobotic rover makes it by 2012, or even 2014, when the Google prize expires, it will arrive years ahead of the astronauts.

It won’t be an easy road. “Skeptics are everywhere,” Whittaker acknowledges. Robert Richards, CEO and founder of the rival Odyssey Moon team, agrees: “As hard a challenge as this is technically, the real hard part is the business plan, and closing that business case.”

So far, outside reactions to Astrobotic’s plan have been drastically different from the reaction Whittaker and Gump got to their LunaCorp proposal more than a decade ago. It’s not just the economics and NASA’s attitude that have changed, says Whittaker. Technology has improved. “There was a time when you actually needed an immense infrastructure and engineering technical capacity to undertake this,” he says. “What’s occurred over those decades is that those tools…have gone from a foggy vision to reality.” He cites a list of developments: Gyroscopes and accelerometers have evolved from complex spinning electromechanisms to “sensors on a chip.” Batteries have doubled their energy density; cameras have vastly improved in resolution, durability, and miniaturization. And, of course, computer hardware and software have advanced steadily.

As for the schedule, if funding is slow to arrive, Whittaker has no problem slipping the 2010 launch date. “Not the kind of thing you get worked up over,” he says flatly. Nor does he worry about the competition. “I definitely don’t lose sleep over other teams,” he says. “That would be a formula for losing.”

Colleagues say that Whittaker is driven. Those who’ve worked with him for years can’t remember when he last took a vacation. But he works on his cattle farm in central Pennsylvania to keep in shape, mentally and physically. “Right now I’m going to the moon,” he says, “and that’s not the thing that’s going to stop me.” So for now, Astrobotic has a man,  a plan, a rover, and a rocket on deck, all waiting for one word: Go.