HOW can you keep a single subsystem failure from turning a billion-dollar spy satellite into orbiting space junk? Break it into a cluster of modules, each of which can easily be replaced, suggests the US Defense Advanced Research Projects Agency (DARPA).
Today’s satellites carry multiple instruments and support systems. This integrated approach minimises weight and launch costs, but increases the chance of expensive delays such as those which clouded the launch of the Pentagon’s SBIRS surveillance satellites. Such complex units can also be disabled by the failure of a single subsystem.
DARPA programme manager Owen Brown envisages splitting complex satellites into a wireless network of modules that communicate with each other across a few kilometres of space. Upgraded or replacement modules could be launched as needed and add themselves to the cluster.
Advertisement
In July, after a round of technology development that included work on wireless networks, fault-tolerant computing, and power transfer in space, DARPA picked a team led by Orbital Sciences of Dulles, Virginia, to design the test system.
The key to making such a “fractionated” system resilient will be the distributed computer systems and the wireless network, David LoBosco of Orbital told the American Institute of Aeronautics and Astronautics last year. So four payload instruments, two data recorders, two high-speed downlinks, and two high-speed number crunchers will be distributed among the modules. If a module is lost it would be relatively easy to replace.
“The key to making the satellite clusters work will be the distributed computer systems”
Indeed, Daniel Hastings of the Massachusetts Institute of Technology talks of the ease of launching new modules and installing upgrades as “an almost seductive advantage” compared with attempting the repair of existing satellites. But while DARPA’s Orbital Express has managed a successful satellite rendezvous, NASA’s DART spacecraft collided with its target satellite.
Initial costs for the fractionated system may be higher because of the need to duplicate support systems but lifetime costs should be lower, says Annalisa Weigel of MIT. Standardised modules “could really knock down the cost of spacecraft, and that’s going to provide phenomenal access to space for entrepreneurs and scientists,” she told 91av.
