Spiral

“In 2004, DARPA hit upon a new way to drive innovation: the DARPA Grand Challenge competitions. These are open competitions where DARPA sets a goal and teams of scientists and engineers from around the country try and tackle it. The first one was to create a car that could navigate a desert racecourse without a driver. First prize was two million dollars. In 2004, nobody succeeded. In 2005, five crossed the finish line. By 2007, the challenge had moved from the desert to city streets, and people are now seriously talking about driverless cars on America’s highways. The lesson was clear: you throw money and talent at a problem, spice it up with a little head-to-head battle, and it’s incredible how fast innovation can happen.

“For their next competition, DARPA went small,” Jake continued. “The Grand MicroChallenge was to develop a robot smaller than a dime that could survive on its own in a woodland environment, without external guidance or power, for a month. The ‘woods’ consisted of a giant terrarium DARPA set up in a storage hangar at Fort Belvoir, near Mount Vernon in Virginia. The first year was a bust. Nobody got close, including our team. All the entrants ran out of juice long before the month was out. But the next year, with Liam Connor’s help, we won it running away.” This got their attention. Liam was a legend-the old man’s wrinkled face was easily the most recognized on campus.

Joe and Dave were nearly ready, and Jake retreated to the edge of the stage. Joe, whose real name was Xinjian, was a classic physics grad student, tall and thin, with wide eyes and a love for detail. He was in his fifth year, finishing his thesis on the mechanics of micro-robot locomotion. He already had an offer for a permanent position in Hong Kong but was holding out for a Millikan Fellowship at Caltech. Gruber was a little more unusual, a third-year muscular fireplug with a flair for public speaking-he’d done some acting as an undergraduate at Yale. Each had their part down pat: Dave handled the audience, and Joe handled the Crawlers.

Joe sat at the microscope set up on the corner of the stage. The scope’s video camera was hooked up to the overhead projector. Joe flipped a switch, and several in the class gasped. A giant creature appeared on the screen: a robotic spider-monster. It scurried to the left, then stopped, turned one hundred and eighty degrees on its six legs, and took off in the opposite direction.

“Say hello to a MicroCrawler,” Dave said. “Arguably the most advanced miniature robot in the world. And don’t worry, it won’t hurt you. The image is magnified a thousand times. It has jointed silicon legs that propel it forward, and a small microprocessor in its body that controls its movement. This one here is our smallest-about the size of a mustard seed. We have models ranging up to the size of a quarter. We build them at the Cornell Nanofabrication Facility using the same patterning, depositing, and sculpting techniques people normally use to make computer chips.”

Jake watched as the students stared at the Crawler, rapt. He smiled at Dave and Joe-the three of them had been through this routine dozens of times, but the spark hadn’t faded. Jake was proud of them. Jake had designed the Crawlers and oversaw the project, but Dave and Joe had done most of the painstaking detail work needed to make the designs a reality. Thousands of hours of struggle, failure, and more struggle. The three of them had gone through years of engineering, design, tweaking, and redesign in creating these little beasts. Doing something that had never been done before was brutally hard, like assembling a model ship in the dark. Make that a ship in a very small bottle in the dark. But they had done it. And-the glory of technology-done once, it could be done again, by anyone, anywhere. All you needed were the fabrication recipes and the right tools.

Joe adjusted the microscope’s optics, zooming out so the students could see the entire petri dish that corralled the tiny robots. Ten Crawlers scurried around while dozens more lay motionless, littering the miniature landscape. One hopped in a tight circle, like a fly with a wounded wing.

“What’s wrong with that one?” a student in front asked.

Joe said, “On this batch, we had a bad liftoff during step twenty. When we put down the piezo actuators.”

“Step twenty?”

“It takes forty-seven separate fabrication steps to make a Crawler,” Dave said. “Forty-nine for ones with a full communications system. It rivals Intel’s most complex computer chip. Five weeks of twelve-hour days to get through the entire process, assuming nothing goes wrong.”