Our research advances the state of the art in modeling and simulation
capabilities to enable high-fidelity, real-time simulation of high-precision
spacecraft formations and clusters through the application of parallel
and distributed technologies.
Modeling and Simulation Algorithms
One of the fundamental issues in the simulation of spacecraft formations
and clusters is that the universe is inherently parallel: the states
and trajectories of objects in space evolve simultaneously. An example
of this is light passing through the telescopes of two spacecraft
and then reflected to a third spacecraft combining those two sources.
Each of the spacecraft executes maneuvers and controls the light
at the same time as the others. Traditional simulation approaches
must compute these effects in order (or serially) for each spacecraft,
so it takes longer to perform the simulation as more spacecraft are
added. This severely restricts the accuracy of the simulation and
the complexity of the models it uses, especially when the simulation
is used in real time to test flight hardware and software.
Finder (artist's rendition)
A better approach is to take advantage of the distributed nature
of the formation and distribute the simulation among multiple processors
in a simulation cluster. This powerful method can produce simulations
that operate in constant time versus the number of spacecraft, but
comes at a cost. We are developing new models and simulation paradigms
in order to permit parallel and distributed simulation of formations.
In addition, traditional simulation techniques are not accurate
enough for this class of missions. Distances in formation missions
can be measured and controlled at the level of nanometers, while
the formation size can range from meters to kilometers. Current real-time
simulation technologies for spacecraft missions typically maintain
relative accuracies on the order of 109, while formation
missions require much higher precision on the order of 1011 or better.
Incorporation of modern integration techniques into the distributed
environment will provide the necessary increase in accuracy.
We are building a new simulation architecture, loosely modeled on
that enables the application and use of the technologies developed
in our research.
The Hierarchical Distributed Reconfigurable Architecture (HYDRA)
is designed for seamless deployment of simulation components and
technologies across a wide array of architectures, from single-CPU
to distributed, mixed- and multi-platform environments. HYDRA automates
the process of connecting and communicating between simulation components,
while allowing these automated behaviors to be overridden as necessary.
One key element of our research is infusion of new technologies
into other active research and flight development programs. HYDRA
and the developed technologies have been successfully infused into
the Formation Algorithms and Simulation Testbed (FAST) as part of
the TPF technology program.
HYDRA is also in consideration for the Space
Interferometery Mission (SIM) (SIM) testbeds.