Mainframe computers have given way to desktop PCs, telephone exchanges to mobile phones, and now traditional satellites are to be challenged by modular spacecraft flying in formation, wirelessly networked, each launched and replaced individually to repair and upgrade the "virtual satellite" in orbit.

That is the goal of the US Defense Advanced Research Projects Agency (DARPA) with its new System F6 technology demonstration programme. F6 stands for "future, flexible, fast, fractionated, free-flying spacecraft united by information exchange".

Programme manager Dr Owen Brown explains: "Future, because this architecture has the potential to change civil and military space systems fundamentally. Fast, because it can be put in place on shorter timescales in response to uncertainty, while flexibility is a panacea for uncertainty".

Fractionated - breaking today's large monolithic satellites down into smaller microsat-like modular elements - is the approach DARPA is pursuing.

"Our brains are hardwired to think of systems within one structure," says Brown. "Treating every element as a wirelessly networked module provides a great deal of design flexibility."

Free-flying is a requirement, to allow the elements to be launched separately and to autonomously navigate and rendezvous in orbit, self-forming wireless data and power networks connecting the physically decoupled elements to create a virtual "spacecraft unified by information exchange" - the "SIX" in F6.

DARPA believes fractionated satellites offer several advantages. One is risk. An expensive monolithic satellite can be lost in a single launch failure, whereas the loss of one module over several cheaper launches could be quickly and easily overcome. Building standard modules rather than one-off satellites would bring manufacturing costs down. Clusters would be adaptable and survivable.


System F6 is one of a number of initiatives the US military is pursuing to make its space forces more operationally responsive. These include building and launching smaller, cheaper satellites more quickly to fill gaps or replace losses. Another is DARPA's own Orbital Express concept for refuelling and reconfiguring satellites in orbit using an autonomous robotic servicing spacecraft.

Boeing's Orbital Express was demonstrated on orbit between April and July 2007, but not without the need for human intervention to overcome anomalies that underlined how challenging autonomous manoeuvring in close proximity in orbit can be. System F6 is even more ambitious.

Individually launching a bunch of microsats and getting them to meet up and work together in orbit is not as simple as calling a meeting and laying on coffee and doughnuts in the conference room. The spacecraft elements have to rendezvous and form a network autonomously, without IT support on call.

"We need a new network approach that treats every box and every subsystem [on the spacecraft] as a uniquely addressable network device," says Brown. Wireless intra-satellite communications will enable computing and other resources to be distributed across the cluster of spacecraft.

If a processor node fails, the other elements can keep the cluster functioning until a replacement spacecraft is launched to join the network. In the same way, a new mission processor can be launched and inserted into the cluster to upgrade the performance of the virtual satellite.

DARPA also wants to demonstrate wireless power transfer between spacecraft - say a powersat with solar arrays topping up a computing node with rechargeable batteries. Candidate technologies include radio-frequency or optical beams, or WiTricity - wireless electricity - which uses the inductive coupling between electromagnetic loop antennas tuned to the same frequency to transfer energy.

"We want to go completely wireless, including wireless power transfer, but the case has to be made for it," says Brown. "In Phase 1 [of the programme], we will look at it and see if it makes sense." And he means wireless within, as well as between, spacecraft.

"We see advantages in going to a wireless spacecraft bus," he says. "In a large monolithic satellite, wiring harnesses can take months to manufacture, and can be 10% of the mass of the spacecraft. They are fraught with quality issues. If we can go wireless in data and in power we can build a spacecraft without wiring harnesses."

Wireless communications and power will also enable a new class of spacecraft component, DARPA believes - a black box-style flight recorder that would allow information to be stored, recovered and analysed rather than lost forever in orbit when a satellite failed for some unknown and undiagnosed reason.

The wirelessly powered recorder would collect data from components within the satellite via Bluetooth-like wireless communications and maintain 90 minutes of spacecraft health and status information. After recovery from the failed satellite, the data could be analysed to diagnose the failure.

A key enabling technology for System F6 is cluster flight. The separately launched spacecraft have to be capable of gathering autonomously and virtually "docking" then flying in formation.

"They can't run into each other, or fly apart," says Brown. "One node knows its absolute position the rest of the nodes know where they are relatively. It's a GPS-like analogue. For the system to make sense, the human can't be in the loop."

The operator will define minimum and maximum spread radii and cluster geometries that the spacecraft will then maintain autonomously. But DARPA also wants to demonstrate the ability to make a rapid, defensive change in cluster geometry, perhaps to avoid an anti-satellite weapon.

System F6 is planned as a five-year, four-phase programme culminating in a year-long on-orbit demonstration of a fractionated satellite made up to two or more modules each weighing less than 300kg. DARPA has awarded contracts to four teams for Phase 1: Boeing ($12.9 million), Lockheed Martin ($5.8 million), Northrop Grumman ($6.2 million) and Orbital Sciences ($13.6 million).

Phase 1 covers developing key technologies designing a system to perform a selected mission developing analytical tools to compare the cost and value of equivalent fractionated and monolithic systems and building a hardware-in-the-loop (HIL) testbed that emulates the fractionated spacecraft using a cluster of networked computers to represent the nodes.

This first phase will include three HIL "block" demonstrations and culminate in a preliminary design review. The second phase will add breadboard wireless communications modules and prototype mission processors and GPS receivers, and will culminate in a critical design review.

"The first two phases are unique in that they focus on building a spacecraft testbed to demonstrate the hardware approach and develop the software in an emulation of the system," says Brown.

"They will build the testbed starting with PCs and using Ethernet. As they spiral in components the testbed gets more fidelity, and in the end it's a complete set of spacecraft elements."

Completion of the individual spacecraft, environmental testing and cross-network integration is planned for Phase 3, which will include an end-to-end ground demonstration leading to a flight readiness review. Phase 4 is the on-orbit demonstration, with the first launch planned within four years of programme start.

The product of the demonstration is planned to be a set of open, non-proprietary interface standards "that read like an IEEE or IP standard, not a specific bolt size". These will enable designers of future space systems to consider using the fractionated approach.

fractions in action

graham warwick washington dc

The satellite of the future could be a cluster of co-operating microsats that can be repaired and upgraded in orbit by launching replacement spacecraft

"Treating every element as a

wirelessly networked module

provides design flexibility"

owen brown

DARPA programme manager

Source: Flight International