Cars to the Cloud
Can Artificial Intelligence Fundamentally Change Formula One?
Artificial intelligence may be the greatest technological disruptor of our time, and sport is no exception. With the Formula One Group partnering with Amazon Web Services to leverage greater machine learning and cloud capability, change is no doubt on the horizon. Callum Tyndall explores this and other ways that Formula One is set to be changed by AI
Announced at the end of June, the Formula One Group’s partnership with Amazon Web Services (AWS) will see the vast majority of the group’s infrastructure transferred from on-premises data centres to AWS’ cloud platform.
The primary advantage of the partnership is that, by standardising with the machine learning and analytics that AWS provides, Formula One can achieve a cloud transformation that will see all of the data that already flows through the sport opened up to new metrics of analysis. In a sport that can be decided by milliseconds, data is becoming an increasingly important tool and the ability to break it down and produce new insights, not just for the teams but also for the fans, is of immense value.
“Leveraging the cornucopia of services offered by the world’s leading cloud, Formula One will engage with its growing global fan base in unique ways,” said Mike Clayville, vice president for worldwide commercial sales at AWS. “Formula One’s years of valuable historical race data analysed against the real-time information that is collected in every race using AWS’s machine learning, streaming, and analytics services will uncover new racing metrics and insights that were unimaginable in the past.
“And, with AWS Elemental Media Services, they will engage their audience through truly differentiated experiences that will thrill generations to come.”
Bringing fans into the decision-making process
One of the core components of the cloud upgrade enabled through the AWS partnership is the use of 65 years’ worth of historical race data to train deep learning models. Aside from simply giving a greater analytical breakdown of how races have evolved and showing how teams can optimise their strategies, these models will also be able to make race predictions and analyse data in real time.
This data will also add value to the fan experience, giving spectators a real-time, data-driven feed of why teams may be adopting a particular strategy and just what is going on underneath the car’s exterior. As the sport looks to make a broader digital push, with a renewed focus on social media and digital platforms, the capability to provide fans with a level of insight that is usually limited to the teams could massively change the experience.
Removing the driver from the equation
Potentially the largest disruption that artificial intelligence and machine learning could bring to Formula One is the advent of driverless vehicles. Even though driverless vehicles are still fairly early into development, making it unlikely that we will see such vehicles capable of Formula One speeds and manoeuvres any time soon, the prospect of autonomous vehicles could transform racing.
Companies such as Roborace are already showing that the concept is viable and that as machine learning develops and more and more data can be fed in (assisted by services such as AWS), cars could eventually become smart enough to fully replace human drivers.
The idea is not without its criticism of course; there are some who believe that in order for Formula One to remain the sport it is it needs the contest between humans and that what makes the sport exciting is the contest of human will and skill. There is certainly some debate as to how you would establish true competition between autonomous vehicles that would presumably be either hopelessly split in their capability or so close as to remove the excitement of content.
However, with the increased safety alone, it is hard not to imagine that we will at some point see driverless cars take to the tracks.
Enhancing the algorithms
In March, Lewis Hamilton took second place in the first race of the Australian Grand Prix after something went wrong with his car’s software. In essence, the Mercedes team’s software told them they had a far more comfortable gap in which to operate than they ended up with. A four-second miscalculation saw Sebastian Vettel take first place.
Mercedes team boss Toto Wolff explained the issue saying: "We were always within this three to four-second margin and then suddenly the cameras showed us the pit exit and Sebastian came out in front of us. The drivers always oscillate in the time delta within one second and he did absolutely nothing wrong. It was down to a software bug or an algorithm that was simply wrong."
While such challenges obviously add tension and excitement to F1 races, they also leave room for improvement by enhanced machine learning. While the algorithm may have told the Mercedes team to expect Vettel to spend 15 seconds in the pit, the new data point received from that race could now be fed into the software to watch out for the Ferrari team’s pit times.
The smarter the software becomes, the more it can start to account for just how a driver can and should react to any given situation. This raises the question of whether we may reach a point where the sport is so data-driven it may as well feature autonomous cars – but for now, there is nothing akin to the thrill of human skill.