American Scientist

By some counts there are nearly 2,500 artificial satellites orbiting the earth. (This tally excludes the countless pieces of man-made debris circling our planet—charred rocket bits, camera-lens caps, that kind of stuff.) Without exception the artificial satellites are relatively large and complex machines. Even the smallest of them, the so-called "microsatellites," weigh tens of kilograms, carry on-board computers and cost millions of dollars to launch. Because they are such complex structures, these costly devices can be disabled by something as paltry as a high-speed grain of dust or some energetic particles from the sun. Such hazards have always made the satellite business a risky endeavor. But all of this may change if a group of scientists at the Los Alamos National Laboratory can get others to consider their revolutionary proposal.

The scientists, Kurt Moore, Janette Frigo and Mark Tilden, presented their creation—the satellite robot, or "satbot"—last December at a meeting of the American Geophysical Union. At first glance a satbot doesn't look terribly impressive: Only a few centimeters across and weighing less than a hundred grams, it looks like something that might be soldered together by a weekend hobbyist. And indeed it can be made almost entirely from off-the-shelf components. What makes the satbot interesting is that it's designed to thrive in an environment that would disable a larger satellite. Not because it's a brutish contraption, but because its inner workings are so simple.

Unlike a typical satellite, the actions of a satbot are not dictated by a software-driven computer. Instead of dutifully following a fixed algorithm, the satbot adapts to its environment by means of a simple nervous network devised by Tilden, a robotics engineer who's created hundreds of different robots (mostly earthbound walking types that live in his office). Tilden's nervous network mimics a biological central pattern generator, much like those that control locomotion in an insect. It's a design that's proven surprisingly robust to noise. "You can almost beat these things into robot roadkill, and the beast will keep dragging itself along," notes Tilden.

One prototype satbot (pictured) has merely two "neurons" in its nervous network. And a neuron itself consists of no more than a resistor, a capacitor and an inverter. "There are no microprocessors in a satbot that can be put out of commission by a magnetic disturbance on the sun," says Tilden.

In fact, the satbot's nervous network is designed so that the little robot's actions are based on electromagnetic changes in its environment. A six-neuron satbot can reliably orient itself to the sun by reacting against the ambient magnetic field, merely using some simple photosensors and wire coils (right). Add an antenna, linking it to a communications satellite, and the satbot could provide information about what's going on in its little corner of space.

Moore, a space scientist who specializes in magnetospheric physics, has been busy devising ways to exploit the satbot's adaptability so that the little robots can be put to work. "Imagine a swarm of hundreds or thousands of satbots launched toward the magnetopause [the boundary between the magnetic fields of the earth and the sun]," he muses. "By monitoring their orientations we could watch the position of the magnetopause as the solar wind pushes it around."

Moore has also been thinking about potential visual applications. "Consider how the compound eye of a fly can provide a high-resolution image even though each photoreceptor provides low-resolution information about light," notes Moore. "With a photosensor each satbot could provide a single 'pixel' of information for a larger image." Arrayed over large areas of the planet, the "pixelsats" could deliver high-resolution images of features on and around the earth.

Aside from adding spice to the Weather Channel, satbots could push the limits of "smaller, faster, cheaper" space technology. Tilden has a vision of producing satbots commercially, a "satellite for the masses." "Imagine a $1,000 satellite with a credit card–sized payload, perhaps one that can carry an imaging device for an amateur astronomer," he says. "Consider the benefits to science if it could be that easy to get into space."

The Los Alamos scientists are approaching NASA and DARPA (the Defense Advanced Research Projects Agency) to fund further experiments with the satbots. "Some of this is still in the 'blue-sky' stage of development," cautions Moore, "but nobody has come out to say it's impossible."—Michael Szpir