Channel surfing - Paul Drouilhet wins Aviation & Aerospace award for MIT's plane navigation without ground control - includes finalists - 1996 Discover Awards

WHAT PAUL DROUILHET LIKES MOST about piloting his own single-engine Bonanza airplane is the feeling of freedom he gets when he's dashing through the clouds--particularly when he's flying over the ocean, where there are no air traffic controllers to dictate what route he has to take. So great is this passion for freedom that he even invented a technology that allows pilots to coordinate their own routes in the air without the aid of air traffic controllers on the ground.

Drouilhet, an electrical engineer at MIT's Lincoln Laboratory, realized that if planes had some way of telling each other what their positions were, accidents would be avoided and costly ground radar rendered unnecessary. Conventional radio communications, however, are too vulnerable to interference for such a safety-critical task. So Drouilhet came up with a new type of radio that is far more robust and reliable. Each plane equipped with his radio first ascertains its own position by listening to the signals of the global positioning system (GPS), a network of satellites by which airplanes have been navigating for years. About twice each second, the radio broadcasts a very brief message announcing the plane's ID, altitude, and position. To keep the messages being broadcast from many different planes from interfering with one another, the radio "squitters" its messages: rather than sending one every half second, it randomly varies the interval between each pair of messages. "The idea," Drouilhet says, "is to have the message so short and the communications channel so wide that there is no problem of interference."

During a test at Logan Airport in Boston, Drouilhet's radio, called GPS Squitter, proved to be exceptionally accurate--it even detected a tiny Cessna being buffeted by a passing commercial airliner. Allied Signal, Rockwell Collins, and other contractors are now working with Drouilhet's group at MIT to develop and manufacture the devices.

FlNALISTS

ORBITAL TINKERTOYS

Los Alamos's Snap-Together Satellite INNOVATOR: TIMOTHY THOMPSON

If the catchphrase of the space business these days is "Small and cheap," then Tim Thompson's motto might just as well be "Less is more." Or perhaps "Looks aren't everything." The new material that he has invented for building satellites is light, strong, and inexpensive, but it's rather boxy, and, truth be told, it makes for a dowdy satellite. "The reason it's cheap," he says almost apologetically, "is that its components are flat. Like plywood, basically."

Thompson is used to being kidded. A mechanical engineer from Los Alamos National Laboratory, he originally designed his material to serve as a particle detector for the Superconducting Supercollider, and when Congress axed that project, he thought satellite builders might appreciate his material's properties. It did not sell itself. "The general impression was 'Hey, this stuff looks like Tinkertoys,'" he says.

In fact, it does. Thompson took a lightweight graphite epoxy composite, the kind commonly used for aircraft parts, and made it stiff and exceedingly strong by gluing it onto an aluminum honeycomb. Then he made standard pieces of various sizes with half-inch-wide tabs and slots on their edges so they could be snapped together, not unlike a child's building toy. The joints are sealed with an epoxy glue gun.

Last summer Los Alamos labs and the contractor Composite Optics of San Diego finished testing the first satellite made of Thompson's materials. It cost 60 percent less to build than a similar satellite made with lightweight composite materials. And although it costs about as much as a satellite made of aluminum, it weighs one-third less--a 48-pound savings that allowed engineers to add an extra instrument. It is scheduled to be launched in October atop a Pegasus XL rocket.

CATCHER IN THE SKY

NASA's Robotic Helpmate INNOVATOR: JON ERICKSON

The only way to field-test Jon Erickson's robot is to take it aboard the Kc-135 aircraft, also known as the vomit comet, while it makes parabolic dives so nauseating that roller coasters seem tame by comparison. That's why Erickson, a chief scientist at NASA's Johnson Space Center, prefers to stay on the ground while his research associates carry out the tests. though he only saw it on videotape, he still swelled with fatherly pride in 1994 when his creation succeeded for the first time in catching a free-floating ball in its three metal-hinged fingers, and again last year when it managed to grab onto a box with a handle.

Grabbing the handle was an especially important milestone because, Erickson says, it is "analogous from a robot's point of view to latching onto a spacecraft." Such a skill is important if Erickson's robot is ever going to accompany astronauts on their flights into space. An astronaut wearing a bulky pressure suit needs an hour or more just to bring tools and parts to a work site in space, for tasks that could be done just as well by a robot helper.

Erickson's robot tracks objects with its stereo cameras and predicts where they're going from their trajectories and orientation. Then its computer instructs the seven joints in its arm and the nine joints in its fingers when and how much to move. The robot isn't really intelligent; it doesn't learn, and it only makes calculations based on data from its sensors and grippers. But it can perform these calculations at the dizzying speed of about 700 million per second. Until now even doing something as simple as catching a ball has required more computing power than robots could muster. "It's essentially a Cray supercomputer inside a robot," Erickson says.