The Gulfstream G550 aircraft. Photo: Gulfstream GD.
SAVANNAH, Ga. – General Dynamics [GD] Gulfstream is developing artificial intelligence and a data concentration network for the G500 as options for future special mission aircraft manufactured here, executives told Defense Daily in an interview on May 9.
The bulk of future development for specific artificial intelligence applications will come from partners such as L3 Technologies [LLL], which is currently modifying four G550s to become the next generation electronic warfare fleet for the Royal Australian Air Force, Leda Chong, vice president of government programs at Gulfstream, said.
But the aircraft with embedded network capabilities that present opportunities for rapid upgradeability and the use of artificial intelligence and machine learning comes from Gulfstream's business jet division, in the form of the G500/600's data concentration network. The new network first entered service on the G500 in 2018, and has not been used yet by Gulfstream's special missions division.
"Artificial intelligence and machine learning have really become ubiquitous in our day-to-day lexicon," Chong said. "In the special missions space we leave that up to the primes that need to integrate the technologies on a per contract basis. We don't go build something because we think some future customer is going to need it. We look for the technical requirements and build to those."
Developed by GE Aviation, the data concentration network uses an Ethernet backbone and can host coding and facilitate computer language translation. It is the first time Gulfstream is using a central network that can perform smart functions and adds new capabilities by reprogramming existing remote data concentrators with software rather than adding new hardware in the form of computers and processors.
For example, when a pilot starts the engines, the network will start the navigation lights and auxiliary power unit fuel pumps simultaneously. On previous Gulfstream aircraft, these functions would have been controlled by separate computers. That presents major weight savings opportunities as Gulfstream continues to modify existing business jets to meet special-mission requirements, Chong said.
"It is a very intelligent network, that enables increased machine-to-machine language translation capability," Chong said. "The network can also present some cost reduction opportunities through weight savings, because the use of that type of network on the aircraft eliminates a significant number of radio racks."
Using the data concentration network on the G500 initially achieved 250 pounds of weight savings, according to Colin Miller, senior vice president of innovation and engineering at Gulfstream.
Around 1,500 engineers at a facility in Savannah dedicated to research and development are evaluating the use of predictive maintenance analytics on critical aircraft parts. Miller describes one of the smart functionalities controlled by the network to include the use of self checking the speed of the opening and closing of the G500's engine valves. The data concentration network tells each valve when to open and close, and an embedded health trend monitoring system measures how fast that is occurring.
"Artificial intelligence and machine learning become important when you want to know how far into an aircraft's life cycle will that valve start to fail," Miller said. "Most of AI is not actually thinking, it is pattern matching. Computers are much better at seeing patterns than humans are. An AI engine can look at how many years a part has been in service, and what're the average number of life cycles it can endure before it will need to be replaced."