The Digital Athlete, from the NFL and Amazon Web Services, aims to prevent injuries from happening

Brendan Hail

The NFL’s collaboration with AWS has helped lead to a reduction in concussions and the creation of position-specific helmets.courtesy of NFL The most important football player in the future of the NFL will do anything without complaint. Run the same play over and over? Absolutely. Wear a new helmet or […]

The NFL’s collaboration with AWS has helped lead to a reduction in concussions and the creation of position-specific helmets.courtesy of NFL

The most important football player in the future of the NFL will do anything without complaint. Run the same play over and over? Absolutely. Wear a new helmet or new cleats? Happy to. Adapt to a change in venue, the weather, even the rules? No problem.

This is the Digital Athlete, a core component of the NFL’s health and safety initiative, which has been developed in conjunction with cloud partner Amazon Web Services. It is a composite, by position group, of an NFL player. Some of it is already functional, some it is still aspirational. All of it is pioneering.

“It’s really a computer simulation model of a football player that can be used to replicate scenarios within the game environment,” said Jennifer Langton, NFL senior vice president of health and innovation. “So, whether that’s variations by position or environmental factors, you can change those factors within a simulation model, and see what the impact could be, which is really, really novel.”

The NFL has been collaborating with AWS on its Next Gen Stats since 2017 and is halfway through an expanded three-year collaboration targeting player safety initiatives. This work is part of the broader $60 million commitment the league made in its Engineering Roadmap. Among the milestones thus far: a 24{c2959109750a408b57b93849e4f6c0b25384840fb4bae13e611646209767c39f} reduction in concussions over the past three seasons (2018-20) compared to the three prior years; a shift from 40{c2959109750a408b57b93849e4f6c0b25384840fb4bae13e611646209767c39f} to 99{c2959109750a408b57b93849e4f6c0b25384840fb4bae13e611646209767c39f} of players wearing top-performing helmets as rated by safety; and the creation of the first position-specific helmet. All of these data points — and much more — helped inform the league’s decision to expand from 16 regular-season games to 17 starting with the 2021 campaign.

The Digital Athlete draws on wide-ranging inputs that generate nearly three terabytes of data per week: video review, equipment scans, mouthguard sensors, game and practice performance data, and much more. This prevalence of sensors, married to video, helps the engineers measure the range of experience for NFL players that can be used to create a digital twin — not an identical Tom Brady or Derrick Henry — but a virtual placeholder replicating the attributes of a player at each position.

“The idea is to create a representation of NFL athletes that we can put into scenarios, and we’re not risking anybody,” said Biocore principal data scientist Sam Huddleston, whose firm is led by NFL Engineering Committee Chairman Jeff Crandall. 

Before joining Biocore, Huddleston spent two decades working for the Department of Defense in various posts for the Army, Navy and NATO.

“We have this phrase that we used to use at the DoD that started out in counter-IED — we’re always working ‘left of boom,’” Huddleston said. “That’s the way I’m trying to approach these injuries: We need to get ‘left of boom.’ So, if we know that something’s a risk and we have time, let’s create time to intervene and get left of the injuries.”

The league had been compiling its NFL Injury Surveillance System for decades, but progress in improving player safety was slow because much of the work was still done manually: reviewing game video frame by frame, extracting injury data and identifying the players. The league only completed about 50 games of such detailed analysis per year.

Four years ago, the NFL went on a road show to meet with technology partners who could amplify the work at speed and scale by building a data lake with a host of inputs ranging from the NFLISS to the Next Gen Stats player tracking data and the head impact registered by its pilot mouthguard sensors program.

Following a pair of crowdsourced competitions to develop efficient algorithms, the same frame-by-frame review was then conducted using the new computer vision techniques. The program completed every game from the past two seasons over the recent Christmas holiday.

“The Digital Athlete was part of the NFL vision when they met with AWS,” said Priya Ponnapalli, senior manager at the Amazon Machine Learning Solutions Lab. 

The Digital Athlete is more of a catch-all term than a specific software application. “To understand athlete injuries, you have to have a comprehensive history of what happened to the athlete before, during, and even to a certain extent after,” Huddleston said. “And, so, we built this program around being able to take every sensor that is collecting information about what’s happening in the game and put it together in an athlete-centered view.”

All of the NFL’s broadcast footage has already been migrated to the AWS cloud, so once the time stamps of the Next Gen Stats and game video are synced, the engineers can retroactively generate data back to 2016.

Video reviews, equipment scans, game and practice performance data and a whole lot more go into the creation of a “digital twin” for NFL players, such as Arizona Cardinals quarterback Kyler Murray.courtesy of NFL

The injury reconstruction work conducted by the team at Biocore necessarily followed an injurious event on the field, but the team never had a comparative dataset. The video analysis combined with player tracking data will fill in those gaps. The engineers have been able to determine that 60{c2959109750a408b57b93849e4f6c0b25384840fb4bae13e611646209767c39f} of all injuries are to a lower extremity, and even identified a closing velocity that makes a player more susceptible to concussions.

The goal is for this work to be done in concert across league departments and committees. At an NFL-sponsored innovation panel last fall, Atlanta Falcons CEO Rich McKay, who serves as Competition Committee chairman, cited the importance of this artificial intelligence and machine learning research.

“We used to project what we thought was going to happen, but we really didn’t know,” McKay said. “We now know. I think that helps us, innovation-wise, move down a path much faster.”

Whether it’s revamping kickoff formations or prohibiting players from lowering their helmet, the engineering team is trying to architect minimally invasive changes.

“We want to preserve the game,” Huddleston said. “But we want to remove the very, very specific things that are injury-causing.”

While a number of tracking and biomechanics analysis options exist for other team sports, they have been conspicuously absent from contact sports such as football and hockey.

Keeping track of which player is which during pileups on the field has always been a major impediment of pose, or skeletal, analysis. Such identification was at the core of two computer vision challenges the NFL and AWS have sponsored, one to automatically tag helmet impacts and the second to ID the players involved.

One way the NFL has begun remedying this, Langton said, is by applying additional computer vision algorithms on top of Intel TrueView’s 360-degree volumetric video to generate skeletal modeling of the players. That is used in tandem with higher-fidelity broadcast footage for a robust 3D motion capture.

Simply knowing what equipment a player is wearing at all times is helpful in case of injury, as is knowing which of the field surfaces are in use, grass or one of 12 unique artificial turf recipes at the 30 NFL stadiums. 

To that end, the NFL installed digital scanners in every stadium and training facility locker room so that each player’s equipment is logged prior to every game and practice. Helmets, shoulder pads and cleats are all adorned with digital tags.

No blade of grass is left unturned. The league is piloting a device that drives across the playing surface before games to map out the football field. It’s like a Mars rover for the NFL and helps fuel a database of interactions between field type and equipment, such as cleats or helmets.

In a side agreement made with the NFLPA, Langton said, all of the players’ wearable performance data is made available to the Engineering Committee — and not anyone directly working at the league office — for the sake of player health and safety initiatives. (The NFLPA declined comment for this story.)

Just as the NFL has made position-specific helmets a priority, Huddleston said the ultimate goal would be true personalization: recommendations of cleats and helmets to each athlete based on game play, venue and weather.

“It really is the first time that we had the comprehensiveness of all of the data that our engineers had access to,” said Langton,” and the Digital Athlete helped us to put it into that environment.” 


SportTechie is now a part of Leaders Group. For more on the intersection of sports and technology, go to

Next Post

Tennis great wants Djokovic answers

John McEnroe says Novak Djokovic should swiftly explain discrepancies discovered in the PCR tests he submitted as evidence to Australia’s Federal Court earlier this year. Pictures: Getty Images Tennis legend John McEnroe has demanded answers from mens World No.1 Novak Djokovic, after reports cast further doubt over the PCR test […]

Subscribe US Now