Scott Zemerick watched a simulated flight of NASA’S SLS (Space Launch System) play out on a TV screen.
He gestured toward the video as it showed the rocket going through the stages of its flight.
The simulation was part of a press conference at the Katherine Johnson IV&V Facility Wednesday to discuss a breakthrough software that will be used on SLS, which is NASA’s replacement for the Space Shuttle.
As the result of a recent NASA Small Business Innovation Research (SBIR) Phase I grant, a team of TMC Technologies of West Virginia (TMC) engineers and scientists developed a new mini-supercomputer and software to ensure space flight software performs as safely and efficiently as possible.
Called OUTLIER, the new software works in concert with a mini-supercomputer called “Gargantua.”
It’s a “Smart Fault Management” system that TMC said “combines a type of artificial intelligence or machine learning, data science and a machine able to perform hyper-fast ‘what-if’ simulations.”
Zemerick talked about its use on SLS, which NASA states will be the “most powerful rocket” it’s ever built and “will enable astronauts to begin their journey to explore destinations far into the solar system.”
“It will be applied to this to help ensure the software does what it’s supposed to do,” Zemerick, a program manager with TMC, said.
He has a lot of confidence in OUTLIER, which he said will start to be applied to SLS in the next six months.
“We’ve done some prototyping and some initial work,” he said. He thought it had a lot of capability.
“OUTLIER is very important because it helps the human analyze lots and lots of flight software data that a person cannot normally analyze by themselves,” he said. “There’s so much data, people can’t do it. It has to be done by computers.”
He said OUTLIER tries to prevent problems — such as a failed sensor — from developing and catch them early.
Dr. Max Spolaor, the TMC team’s Chief Scientist and Principal Investigator, noted the significance of OUTLIER.
“I believe it’s quite a breakthrough in the field,” he said. “We call it Smart Fault Management because we’ve been trying to use these techniques and concepts from emerging technologies, so it’s something new that’s not been done before and it’s proven to be a very robust tool and very useful.”
According to TMC, fault management is “a process of detecting, isolating and ultimately resolving various software or network problems.”
Spolaor noted a lot of data is generated.
“And when we generate these data, we generate also these failure scenarios,” he said. “Then, we can analyze how the flight software reacts to these things, and we can prevent them from happening.”
“We want to prevent and mitigate all these failures and anomalies and try to catch them now rather than in space when the thing is flying.”
He thought it was applicable to other areas such as cybersecurity and homeland security.
M.J. Durst of Blacksville, a senior at WVU and an intern with TMC, played an instrumental role with “Gargantua.”
“I was kind of a starting point to the research where I kind of did very basic concepts, and then showed it to them to make sure it worked and then from there I kind of just optimized the software as it came out,” he said.
Zemerick states that Gargantua consists of “a series of highly advanced, Graphic Processing Unit (GPU) workstations specifically designed to handle the demanding processing and intensive scientific and engineering applications OUTLIER will need to perform on the spaceflight software.”
“The TMC-built Gargantua has a calculated, theoretical speed benchmark of 16 teraFLOPS,”
Zemerick said in a press release. “In simpler terms, our mini-supercomputer can perform 16 trillion operations per second.”
TMC is described as a West Virginia-based information technology services company focusing on the space, defense, justice, cybersecurity, energy and natural resources/climate sectors. NASA is one of its customers.