If aerospace research is still a bellwether for commercial developments, expect more miniaturization. A hint of things to come was presented at the American Geophysical Union's semiannual meeting by researcher Janette Frigo and robotics engineer Mark Tilden of Los Alamos National Laboratory, Los Alamos, N.M. They envision robust microsatellites--down to a half ounce--as one evolutionary trend in space exploration. One possibility: using clusters of small, cheap satellites that can achieve results even when some of them fail. The approach grew out of Tilden's work with a simplified approach to the control of experimental, insect-like robots. Based on simple "nervous-nets" that work like neurons in animal nervous systems, the control system for the satellites would have no microprocessors or fixed algorithmic behaviors. A microsatellite controller was demonstrated at the meeting.