Science

College athletes build AI-powered wearable designed to prevent sports injuries and expedite healing


College athletes build AI-powered wearable designed that can predict injuries before they happen and speed healing

  • Fathom AI relies on three small sensors to collect over 100 biometric readings 
  • The device will emphasize downtime and recovery and minimize overtraining
  • The product is still in early testing but is planned to retail for $299 

A new wearable app and sensor promises to help professional athletes minimize their risk of injury.

Developed by two former student athletes at Duke University, the app will use biodata from athletes to customize training routines for individual athlete in ways meant to minimize injury and prevent overtraining.

Called Fathom AI, the device consists of a smartphone app and three sticker-sized biosensors that attach to a person’s ankles and lower back and collects readings on more than 100 different biodata categories.

One of three small sensors used by Fathom AI to collect biodata from athletes and develop custom training advice and guidelines to minimize the risk of injury.

One of three small sensors used by Fathom AI to collect biodata from athletes and develop custom training advice and guidelines to minimize the risk of injury.

‘Many people don’t realize that the most effective injury prevention may start with just 10 or 15 minutes daily of data-driven prep and recovery to keep small things from becoming serious injuries,’ co-founder Ivonna Dumanyan told WRAL’s TechWire.

The sensors track fatigue, variation in performance between days and between exercises among many other metrics.

The metrics are analyzed by an AI to determine whether an athlete might be overtraining or pushing their body into states where injury could become more likely.

The tech can also work in concert with other biometric data apps like Apple Health or Garmin’s smart watches.

An AI-driven app will organize and interpret Fathom AI's biometric data and customize advice for each individual athlete.

An AI-driven app will organize and interpret Fathom AI’s biometric data and customize advice for each individual athlete. 

Fathom AI was founded by Gabrielle Levac (left) and Ivonna Dumanyan (right), two former student athletes at Duke University.

Fathom AI was founded by Gabrielle Levac (left) and Ivonna Dumanyan (right), two former student athletes at Duke University.

HOW DO FITNESS TRACKERS WORK? 

Fitness trackers such as Fitbits or smart watches monitor heart rate using a technique called photoplethysmography.

The tracker sends green light through the skin which is partially absorbed by arteries.

As you exercise, these arteries expand as blood flow increases – meaning more green light is absorbed rather than reflected back to the tracker.

The tracker estimates your heart rate by seeing how much light is reflected back. 

The amount of light that passes back through the skin to the tracker can be affected by the amount of melanin in the skin, and any tattoos.

‘We essentially build this massive model on how you move that is unique to you,’ Dumanyan said.

‘From that, we create a hyper- personalized exercise plan to balance the stresses on your body and correct those imbalances.’

The devices are still in limited testing, but potential users can request access through the company’s website.

The plan is to eventually launch the sensors as a full commercial release, priced at $299 plus a $29 monthly subscription.

Fitness trackers have been popular in recent years, but it’s still unclear how accurate and effective the technology used in them is.

Last year, a survey of Fitbit monitors found they recorded an accurate count of steps a person took in a day at only around 50 percent of the time.

Garmin’s Vivosmart 4 generated headlines earlier this year when researchers discovered some of its units gave inaccurate distance readouts, requiring runners to travel as far as 37 miles before giving them credit for having done a 26.2-mile marathon.

In 2017, a man reported having his life saved by his Apple Watch, when it alerted him to a sudden elevation in heart rate that was brought on by a blood clot, something that could have killed him without emergency medical treatment.

‘Never thought a stupid lil’ wrist computer I bought two years ago would save my life,’ Green tweeted.

‘Saw my heart rate go up, ended up being a pulmonary embolism’

WHY FITNESS TRACKERS AND SMART WATCHES CAN BE INACCURATE

There are two ways that a fitness watch can measure the distance a person has ran, walked or jogged.

Some step trackers don’t have a built-in GPS – so will calculate distance travelled on the number of steps that have been taken.

This is based on number of steps multiplied by average stride length – which can affect accuracy.

Other watches use GPS models or connect to your smartphone to use GPS.

GPS models are more accurate – but Which? results showed some that overestimated distance travelled by up to 20 per cent.  

How was accuracy measured?  

Tests often include having participants walked on a treadmill at 4.8km (3 miles) and hour for 10 minutes and run at 9.0-10km (5.6 – 6.2 miles) per hour for 10 minutes for Which?’s tests.

They are also put through a series of daily tasks, such as loading and unloading the dishwasher and carrying shopping.

To test calorie’s burned, testers wore a face mask that was linked to a gas-analysis system. 

This measured oxygen intake and carbon dioxide production for each breath, meaning calories were measured as accurately as possible.

Tracker results were then compared with gas-analysis data to find the most accurate devices. 



READ SOURCE

Leave a Reply

This website uses cookies. By continuing to use this site, you accept our use of cookies.