How Williams Martini Racing will use machine learning and the Internet of Things
http://www.vbprofiles.com/tb/topic/52a0 ... -of-things
Artificial intelligence, machine learning and the Internet of Things could all play a significant role within the Williams Martini Racing Formula 1 team in the not too distant future, its IT director, Graeme Hackland, has told Computing.
Formula 1 often sits at the cutting edge of information technology, with analytics and big data now playing a key role at the pinnacle of motorsport.
"There's more than just the car, we've almost instrumented it completely and if there was unlimited size for black boxes on the car, maybe we'd be gathering even more data. I think we've got as much as we can get off the car," said Hackland, who prior to joining Williams in January 2014 served as CIO of the Lotus F1 team.
However, the connected devices and other innovative technologies could aid Williams away from the race track in other vital areas including manufacturing and car design. Indeed, rival F1 team Infiniti Red Bull racing are also examining how connected devices can benefit the sport.
"I think where the Internet of Things is going to play a big part is putting sensors in all sorts of areas: the factory, manufacturing, the design side of things. I think we can see improvements to traditional CAD drawings using sensors and technology in manufacturing," said Hackland.
He also explained how machine-learning algorithms could play a crucial role in the fast moving world of Formula 1. "A Formula 1 car is in constant evolution. Although we only make four or five cars a year, every single race that car changes. There are times when you'll manufacture something while the designer is redesigning it," he said.
It's here that machine learning and artificial intelligence could play a role by causing the machine manufacturing the part or the car to automatically react if there's a change made to an associated design file.
"We think with machine learning, AI and the Internet of Things, if we can join up that lifecycle so if a designer opens a part and starts to amend it, the manufacturing capability will know that's the next job," Hackland said.
"If a designer is redesigning it the likelihood is you're not going to use the original part, the machines can start to make decisions about what they're manufacturing," he added.
And that might just be scraping the surface of how the Formula 1 team will eventually harness the power of machine learning.
"In the F1 lifecycle of the car, we think machine learning is going to benefit how we manufacture the car," said Hackland, who offered insight into some ways algorithms and automation could benefit Williams Martini Racing.
"The decision-making that's made on a daily basis around what you put into the machine shop, the optimum way of designing next year's car, is fairly manual now. We're automating, but machine learning will help us with how we manufacture the car.
"There are lots of areas we can apply machine learning that will feed into Formula 1 and make us more competitive," he added.
When it comes to the technology companies providing the software, Hackland is less interested in the size of the vendor than what they can offer.
"I'm trying to make sure we don't only talk to the established companies that everyone knows but also start-ups," he said, describing how the team's partnership with BT helps in this area.
"Working with people like BT, who go out to Silicon Valley and talk to the start-ups, they can give us access to technology we'd have difficulty finding," Hackland said, adding: "I'm keen not to just use established technology."
Ultimately, if using innovative new technology can help Williams get to pole position, the team wants to use it.
"If it's the reliability of our car, we're not going to take any chances. But why not try bleeding-edge technology in other parts like manufacturing and design and get a jump on the other teams before they use it," said Hackland
