New NSF grant allows software engineer Robyn Lutz to investigate the
safety analysis of evolving product lines.
Change is inevitable.
But change can also be risky - very risky when the change occurs in safety-critical
systems.
"Changing a product line of safety-critical systems can be sometimes
be even riskier," says Robyn Lutz, professor of computer science. "But
since change is inevitable, we should seek better ways, such as analysis
techniques and software tools, to reduce the associated risks."
Lutz, who as a software engineer has studied product safety as it relates
to computer technology, has recently received a three-year, $300,000 National
Science Foundation (NSF) grant to continue to look at the safety analysis
of evolving product lines.
When change occurs in a product line, the safety consequences can be even
greater because multiple systems are involved.
An example of such a change is a pacemaker/defibrillator. The software in
past pacemakers couldn't tell when the individual was resting or exercising.
New versions of the product have added a "rate-responsive" software
feature to distinguish whether or not the user is currently active and to
use this additional information in determining whether or not to stimulate
the heart.
Sounds good, right?
"The capability to do this did not exist and had not been anticipated
when the product line was initially built," Lutz says. "But it
had significant safety implications when added as an optional feature to
the product line later on.
"However a business can't sit still and not improve their product,"
she continued. "They have to add more features to keep their share
of the market. We're trying to figure out how best to we keep that product
evolution safe (for the consumer)."
Lutz's research is investigating types of evolution that are both common
and problematic in safety-critical product lines while extending safety
analysis techniques to efficiently update the product-line safety analysis
as the software evolves.
Her research team is also developing automated tool support.
"The number of safety-critical product lines on the market is expanding
rapidly," Lutz said. "We need better analysis techniques to determine
whether a new feature can be added to all the systems in the product line
safely."
Some other of these product lines include aircraft flight-instrumentation
displays, communication satellites, medical-imaging systems, smart vehicles
and assistive robots.
Lutz's area of research is software engineering, the part of computer science
that studies how to develop better software systems. Safety-critical software
is the software that can cause, or that can prevent, a hazardous situation
or loss of a project.
"I investigate methods by which we can more accurately describe and
more rigorously analyze the requirements and design for safety-critical
software and build tools to help developers apply these methods," she
said.
By developing these tools, Lutz hopes these safety-critical issues can be
found earlier in the product development.
"What I would like to do is provide software tools and techniques that
allow companies building high integrity product lines to use these tools
efficiently and cost effectively to check to see if the changes they are
making are good changes," she said.
"If we find these safety concerns earlier in the product development,
it will save a lot of money. If we on't find these concerns until the product
is being tested, it is much more expensive to fix.".
Robyn Lutz
Around LAS
October 30 to November 12, 2006
