With billions spent building, launching—and sometimes losing—geostationary (GEO) satellites, Pentagon researchers have begun a program to cut the cost, and risk, by reusing elements of retired spacecraft already in orbit.

With the U.S. Naval Research Laboratory (NRL) acting as system integrator, the Defense Advanced Research Projects Agency (Darpa) has assembled a broad-based industry team for its Phoenix program to demonstrate the salvage and reuse of components, such as large antennas, from dead satellites. An in-orbit test is planned in 2015-16.

“GEO satellites have an end-of-life because of running out of fuel, or solar panel degradation, or transponder technology becoming obsolete, but usually the antenna is perfectly OK,” says Seamus Tuohy, director of space systems for team member Draper Laboratory. “The concept is to scavenge a large aperture off a cooperative retired satellite, place it in GEO and construct the other satellite functions around it.”

Large antennas drive the size of satellites, and in turn rocket boosters, because there is a limit to how much they can be folded for launch. “The rule of thumb is 3:1. The largest launch vehicle available limits the largest size of aperture that can be deployed—the rest of the satellite can be pretty small,” says Tuohy. “Phoenix takes the driving determinant of satellite cost—the aperture—and uses it beyond its original purpose.”

In the Phoenix concept, payload orbital delivery systems (PODS) containing small “satlets” performing the other functions of a satellite will be launched as ride-share payloads on GEO launches. The PODS will be ejected upon reaching orbit and a tender spacecraft with a robotic arm will capture and store them on an internal tool belt before traveling to the geostationary-satellite “graveyard” orbit.

Each satlet will have a common shape to fit the same payload ejector and tool belt, but a different function. In graveyard orbit, the tender will rendezvous with a retired spacecraft and use its robotic arms to attach the satlets stored in the PODS to the back of the old antenna and aggregate the functions of a new satellite. The arms will then detach the antenna with satlets attached, and the tender will move the assembly to a new operating location in GEO.

Under the Phoenix program, ATK is to modify an existing U.S. government-owned geostationary satellite bus for the demonstration mission. The spacecraft, which will be delivered to NRL by October 2014 for integration and test, is designed to be capable of supporting, for a minimum of a year, robotic rendezvous and proximity operations, and the grapple-and-repair demo.

ATK, in partnership with the University of Maryland's Space Systems Laboratory (SSL), also will develop robotic servicing tools and software to enable reuse of the antenna and other working components of a nonfunctional satellite. Part of the work will leverage technology now being developed by ViviSat, a joint venture between ATK and U.S. Space. ViviSat is working on a free-flying satellite “jetpack” called the Mission Extension Vehicle (MEV) to extend the service life of satellites that have run out of station-keeping propellant. The MEV is designed to robotically dock with a dead satellite by extending a device into the throat of the spacecraft's kick motor, and then latching on.

Space Systems/Loral will study how small satellites can be carried into geostationary orbit as hosted payloads on commercial spacecraft. Communications-satellite operator Intelsat will study hosted payload interfaces. NovaWurks of Los Alamitos, Calif., will design the PODS.

Pasedena, Calif.-based MDA Information Systems is under contract to develop a robotic arm for the ATK-provided tender based on the seven-degrees-of-freedom dexterous arm it developed for Darpa's Front-end Robotics Enabling Near-term Demonstration. New York-based Honeybee Robotics Spacecraft Mechanisms will develop two types of new telerobotic end-effectors. Cambridge, Mass.-based Energid Technologies will provide control and simulation software for the robotic arms. Louisville, Colo.-based Altius Space Machines is building an extendable boom for the tender, and NASA's Jet Propulsion Laboratory (JPL) will provide adhesive pads called “Gecko-Grippers.”

Aurora Flight Sciences is responsible for design and integration of the satlets and testing of prototypes. The Massachusetts Institute of Technology will provide control design expertise and microthruster technology enabling the satlets to point the antenna. JPL is responsible for software development and testing.

Draper will provide attitude determination and control for the reconstituted satellite, using a mix of hardware and software. “It could be more than one satlet,” says Tuohy. “We are looking primarily at software, using techniques developed for the ISS [International Space Station] to optimize control of slewing the satellite.” The zero-propellant maneuver technique uses the atmospheric drag differential of the solar-array sails to maneuver the ISS.

During the 18-month first phase, Draper will produce a preliminary design for an attitude determination and control system architecture that will determine how functions such as star trackers, thrusters and other units should be packaged in the satlets, Tuohy says. A key goal for Draper's work is to take greater advantage of momentum to control the vehicle rather than using thrusters

“Fundamentally it is about optimizing the management of momentum,” Tuohy says. “One of the limits of using satlets attached to an antenna is the use of thrusters to control attitude and dump momentum. That limits the life of the reconstructed satellite. The more optimal the control of momentum, the more life you get.

“In GEO the main external torque on the satellite is the solar wind. Instead of fighting it or living with it, we use it for control,” Tuohy says. “Instead of going from point A to point B by the traditional shortest path, we define a path that minimizes the buildup of momentum so that it does not have to be dumped later.”

Draper is also working on distributed control. “The positioning of the satlets on the back of the antenna is not predetermined. We have to automatically characterize the satellite as more satlets are added and its mass increases,” Tuohy says. “We have to learn how the aggregrated spacecraft maneuvers, as we have done with the ISS as they have added and removed elements.”

Rather than invite bids from teams lead by large system integrators, Darpa selected individual entities for each element of the Phoenix program to assemble a team under NRL. “Darpa is riding herd, for now,” says Tuohy. “The intent is for performers in the individual technical areas to come together and work as a team.” Instead of a “best team,” he says, Darpa selected the best contributors “and hopefully we will make a team out of it. They are trying to get the best of the best versus a large integrator.”