The U.S. Defense Advanced Research Projects Agency is seeking a third party willing to make available a spacecraft to participate in the planned 2013 on-orbit demonstration of a fractionated satellite network.

Darpa's System F6 program is developing the capability to network a cluster of free-flying satellites so they can wirelessly share resources such as processing, data storage, sensors, communications relay and navigation to replicate the capability of a single larger spacecraft.

Prime contractor Orbital Sciences plans to build three modular spacecraft for the demonstration, but the agency is looking for a fourth, non-Darpa satellite that could be equipped to link into the System F6 network to test the interfaces being developed to allow spacecraft to share resources in real time.

The third-party satellite would host an "F6 tech package" enabling it to communicate with the cluster of Darpa spacecraft. The demonstration would involve simulating the failure of certain spacecraft systems, such as a star tracker for attitude determination, and replacing the lost capability by sharing access to systems on other satellites.

Orbital received a $47.6-million contract last month for the year-long Phase 2 of System F6, which will take the program through the critical design review. "An explicit objective of this phase is to identify a non-Darpa user . . . that will link with us to do the demonstration," says Paul Eremenko, F6 program manager.

To that end, he says, Darpa has asked Orbital to modify its Pleiades architecture for System F6, which is based on a cluster of modular spacecraft flying in loose formation kilometers apart on low Earth orbit (LEO), to include the ability to extend the wireless resource-sharing links to satellites in geosynchronous orbit (GEO) and to add multi-level security.

The original concept behind the fractionated satellite was to deconstruct a monolithic satellite into modules that could be launched separately and maneuver autonomously to rendezvous in orbit, linking up to perform the mission of a large spacecraft. Individual modules could then be swapped out to repair or upgrade the constellation on orbit.

Eremenko says System F6 has evolved into an infrastructure in orbit that would allow resource sharing within a cluster, between clusters and with other individual satellites in LEO or GEO. "The program vision is to create a flexible infrastructure able to support as many payloads as possible," he says.

The two key new technologies in System F6 include multi-body formation flight -- not just the ability to avoid collisions between three or more free-flying spacecraft, but also the capability to perform a defensive scatter maneuver on command to avoid debris, or a threat, and then reform the cluster autonomously.

The second is to form a packet-switched Internet Protocol network on orbit, creating a "virtual mission bus" that wirelessly links the spacecraft to allow real-time resource sharing. "We want to be able to replace the control processor virtually, and have another module take over the inner-loop control of that spacecraft, or replace the star tracker with relative navigation from another module," Eremenko says.

Orbital is evolving the architecture through a series of hardware-in-the-loop iterations, beginning with software-only simulations and progressively substituting first breadboard equipment then engineering development units until the test bench morphs into the flight unit. Phase 1 took the program through the preliminary design review and included three software-in-the-loop spirals.

Phase 2 will involve two major iterations, but internally the program is running on a two-week build cycle because "IBM is providing the virtual mission bus, and that's the way they do things," Eremenko says. "This is a software information-technology program and they're in the driver's seat." Phase 3 will take the hardware from engineering development to flight units. Phase 4 is the on-orbit demonstration.

Darpa plans three modules. Two will supply data-relay services. Of the two, one will provide X-band direct to the ground station when in view, the other will uplink to a GEO relay satellite to allow a continuous communication capability at lower bandwidth. The third would provide computing and data storage for the cluster. "All will have some sort of relative navigation," he says.

While the LEO modules will use carrier-phase differential GPS for navigation, availability of GPS signals in geosynchronous orbit is poor, so two ways to provide navigation resources to a GEO satellite are being looked at. One is to triangulate from the cluster using the difference in arrival times of signals from each module. The other is to use a dedicated relative-navigation sensor, either laser or radio-frequency.

Design of the spacecraft bus is still open, says Eremenko, although the plan to make the modules compatible with the secondary payload adaptor ring on the U.S. Air Force's Evolved Expendable Launch Vehicles limits their weight to 180 kg. (400 lb.) A launch has not yet been manifested and the modules could be orbited together on an EELV or separately on two Orbital Minotaur or SpaceX Falcon boosters -- the latter offering a more convincing demonstration of the cluster's ability to rendezvous autonomously.

A 6-12-month demonstration is slated; much depends on the third-party satellite. Darpa is looking for a user planning to fly a payload with a mission relevant to the warfighter and needs to have the third party on board by late 2010.