Ever stop and think about the millions of dollars spent on fancy
space equipment that breaks down once it becomes airborne? If you are
millions of miles away orbiting the Earth, there’s no repairman
available to fix the problem.

Researchers at the University of Michigan’s College of Engineering
think they may have the answer: machines that are smart enough to
learn from experience, detect problems and fix themselves. Atlantis
docked at the International Space Station in September, carrying MACE
II, an on-orbit demonstration of self-reliant, adaptive control
technologies, which is the first scientific experiment aboard the
space station. This week, MACE II begins its mission: to detect
problems with the spacecraft’s hardware and onboard technologies and
correct those problems as they arise.

“This new technology would allow the space station and other space
systems to operate as if they had a human operator, taking note of any
glitches and fixing those problems on the fly,” said Professor David
Hyland, chair of Michigan’s Aerospace Engineering department. Hyland
developed the self-reliant, adaptive control technology used in the
MACE II project. “Problems can be fixed on the spot, so unexpected
hardware failures won’t threaten the life of a space mission,” he
said. “The assistance of ground control will only be required to
perform tasks at the very highest level.”

This technology will create significant cost savings for future
space missions because it reduces the need for an extensive ground
control staff to track and repair errors that occur in the hardware
and onboard technologies.

“Building spacecraft that are more autonomous and reducing the
human factor now required for spacecraft management will inherently
result in lower costs and permit more flexible and reliable missions,”
said Rory Ninneman of Air Force Research Laboratory’s (AFRL) Space
Vehicles Directorate. The AFRL is leading the first science team on
the MACE II project.

How It Works

MACE II relies on algorithms that enable it to adapt to changing
conditions and correct problems without using a ground controller to
exchange messages with the space shuttle. This technology, known as
frequency domain expert control, is an advance beyond MACE I, which
was able to test and fixed gained controls but unable to adapt to
unforeseen changes or to detect faults and failures in the hardware
and technology.

“Since communications in space aren’t instantaneous, we wanted to
develop a technology that would eliminate the need for directions from
ground control,” said Hyland. “The improvement made in MACE II
accomplish that goal and allow us to fix problems quickly. For
example, if one of the actuators of an onboard control system fails,
you don’t want to waste a lot of time waiting for ground control to
explain how to fix the problem. The technology we are trying to
demonstrate on MACE II can fix the problem without human
intervention.”

Experimental Testing at Ground Zero

Experimental testing of MACE II will begin this week now that
astronauts have attended to their many Space Station readiness and
checkout tasks. A variety of different kinds of tests will be
performed using the MACE hardware which, originally developed by MIT,
is a scaled model of a large space platform containing
precision-pointing antennas and control actuators.

“Although we conducted several experiments in similar conditions
on the ground, we are excited to see how MACE II will perform in
ground zero conditions,” Hyland said. “There are subtle change from
one ”g“ to zero ”g“, like rattling of the structure’s joints, so we
will test the algorithms that we have written to make sure everything
works once it’s up in space.”

Once the experiments are completed and algorithms for the
technology are finalized, researchers hope to use MACE II in future
space missions conducted by NASA and the US Air Force.

Who Is Involved?

There are two science teams involved in the MACE II project. The
first science team is led by the Air Force Research Laboratory (AFRL),
and the second team is led by MIT. The AFRL team is a collaborative
project between the University of Michigan (UM), Planning Systems,
Inc., Melbourne Controls Group, Payload Systems, Inc., Virginia Tech
and Sheet Dynamics Ltd. The project was funded by a Small Business
Innovative Research (SBIR) contract, an Educational Partnership
Agreement and a Cooperative Research and Development Agreement.

Researchers at The UM Department of Aerospace Engineering have
been working with the self-reliant, adaptive control technology since
the SBIR Phase II Grant was awarded in 1998.

About the University of Michigan College of Engineering

The University of Michigan College of Engineering is ranked among
the nation’s top five engineering schools according to a recent U.S.
News and World Report study. The College’s graduate programs are
ranked number four, and its undergraduate programs are ranked fifth.
The aerospace department is ranked fourth, and its undergraduate
program is ranked second in the nation.