To answer all those who wonder why engineers would ever need nanocircuits and computers, researchers at Mitre in McLean, Va. (www.mitre.org), are building the world's smallest walking robot. Measuring 5mm from front to back, the six legged robot will be untethered and autonomous, carrying its own power supply and control hardware and computers.

The design team decided on six legs to ensure the robot would have a statically stable gait, which gives it accurate motion control over uneven terrain. Such a gait lets some of the legs move while the remaining legs support the robot. So three legs must stay on the ground, and the center of mass must stay within the triangle defined by those three legs. If there are only four legs total, this becomes a daunting control problem and the design must include a moving counterweight. But with six legs, three legs can be moving while the remaining three keep it fully supported.

Each leg is identical with three degrees of freedom, which lets the robot move in straight or curved paths over uneven terrain. Making the legs all the same simplifies fabrication and control. And placing them on a bilaterally symmetric body lets the robot travel forward or backward using the same gait. The legs use only motors and drive mechanisms already fabricated and proven in other MEMS devices.

For power, the team chose solar or photovoltaic and a battery. This eliminates the need to carry bulky fuel and lets the robot operate in dim light or darkness and leaves open the option of using both battery and solar cell if more power is needed. Batteries as small as 50 m m square are being built by Bipolar Technologies, Provo, Utah. They are fabricating a 1-mm 2 version with a 7-V output for Mitre's robot.

Current plans call for using seven nanocomputers mounted on the robot's back for control. Once the robot is built, researchers plan to shrink it even further, going down into the submillimeter domain.