Learn about Marianna’s work and how she uses AI and machine learning to increase athletic longevity with wearable robotics. One of our 50 Women in AI.
Wearable robotics for fitness and longevity
We work on wearable robotics products that help people to measure the quality of their movements to learn sports, to learn the best way to exercise, avoid injuries and to build their own muscle mass and joint flexibility, also after the age of 45! Because our aim is to make people stay fit for longer, so our focus is on athletic longevity.~ Marianna Recchia
I just returned from a long holiday weekend in the Alps where, between the middle aged-people and ancient dogs, there were lots of complaints about creaks and pains, resulting in an activities list that skewed more towards ingesting calories rather than burning them. No doubt other groups seeking alpine air that weekend were less creaky and more fluid. Or maybe they just decided to push through the various performance glitches and keep moving. But whether opting for naps, beers or peaks, many would like to perform better. Tonus Tech may have a way to help us do just that.
Background: Business designer, innovator, automotive industry R&D
Tonus Tech is a UK based AI powered robotics startup headed by CEO Marianna Recchia. Marianna is not a coder, software engineer or data analyst. Instead Marianna is an entrepreneur and innovator, bringing her industrial background in automotive automation, car telematics, and R&D to the world of robotics for wellness. She made the decision to leave corporate multinationals behind to pursue a project more in alignment with her values of making the world a better place – a new fulfilling challenge where she’s putting her innovation background to good use.
I have known Marianna for many years. Her ever-present, contagious enthusiasm shined through as she talked about building a startup that is working to make our lives better, especially those of us who are no longer blessed with the glow and recovery capabilities of youth and want to remain active and injury free, continuing to enjoy the activities we love.
At the time of our call, Marianna was in Berlin, dividing her time between there and London, after having returned to Berlin after a year at Odense Robotics Accelerator in Denmark. In between her northern European life she tries to return to her native Puglia to visit with family and grab some sun. Our call was a chance to catch up and for me to learn more about her journey with Tonus Tech, how they are applying AI and the challenges they are facing.
When discussing some of Tonus’s challenges, a determined yet practical mood came through the zoom screen. Although she quickly return to her beaming, dimpled grin as we talked about all of the cool possibilities that these wearable robotics provide for those seeking to extend their active lives (and improve performance!).
Below some highlights from our interview.
Tonus Tech’s current product is a suit. The suit contains two different kinds of sensors that provide in-depth info on joint and muscle activity. The data is collected from these sensors and AI and machine learning is used to classify and score the movements. Through machine learning, the person’s quality of movements are analyzed in comparison to optimal movements, showing the gap from where they are to where they should be, in order to be more efficient and reduce chance of injury.
Tonus is also investigating the use of soft robotics technology that provides active feedback or “guides” the wearer’s movements, in order to improve their movement quality and performance.
AI and Machine Learning to reduce injuries
Through machine learning analysis, the Tonus team is looking for injury risk factors, and while professional athletes have benefited from detailed examinations of their movement patterns, amateur athletes and weekend warriors have not had such opportunities.
So for example hamstring injuries, are not just a problem of football players, but (they are a large problem) for people in their everyday life, especially to people in their 50s. They are a relevant problem for women, so for us it ‘s very important to find this injury risk. And also some conditions like sarcopenia ( age related, involuntary loss of skeletal muscle mass and strength.) that people can start to develop already in mid life not just in their 70s. This is where ML could be really relevant.
Collecting data – what kind and from where?
Along with CEO Marianna, Tonus’s team includes a medical doctor, a robotics engineer, a software engineer and a data scientist. With a heavy focus on gathering the needed data, Summer 2022 features a number of product tests. Putting their system to the test in biomechanic labs, they will also compare the data generated by their robotic suit system, to data that is captured from external sources such as cameras, comparing the value and quality of data across the of diverse sources.
Data collection is key for their products efficacy. And it is a challenge to gather the vast amounts of data required for the AI algorithms to successfully identify patterns and generate useful information.
One way Tonus Tech is seeking the needed data, is by working with universities that specialize in biomechanics and sports science. University of Newcastle is one such institution, helping Tonus to understand what kinds of data they need. Marianna has cultivated a highly collaborative approach, building relationships with organizations, including gyms, who have the needed knowledge and user base to first help them identify the type of data that would be useful ( for example data that identified illness or injury patterns) and then figure out how to best capture it.
The challenges of data collection
Additional barriers to data collection include privacy, especially in the health care sector. Tonus works with data from diverse sources, but some information simply doesn’t exist yet. Pushing into new territory with their wearable robotics suit, they need data that helps them to analyze movement, so their trials will help them to build up their data base to train their model. Marianna plans to supplement these trials with grant applications that could allow them to access larger biomechanics data base sets.
Marianna noted that they are not alone in confronting the challenge of collecting needed data, it is a problem shared by many companies that do not already have a large existing database for the information they need. And as mentioned, some data simply does not exist yet. She comments on revised expectations in these cases: Some kind of models are also new so in these cases we we are not expecting a kind of perfect intelligence.
In regards to the challenges she’s facing to find the data needed to help reduce risk of injury:
It’s not easy yet to prevent injuries, so the analysis of these risk injury factors is a huge challenge for any company, both in the field of orthopedics or sports science, and so any company that is currently trying to prevent injuries is dealing with the same problem. It’s still a kind of r & d phase for all models that need to do this.
On the positive side, Marianna feels that this challenge of data collection will lead to new collaborative business models that will help all companies working in health and sports tech, as well as broader collaborations in general.
A longer active life for all of us
At the time of writing this, 36 year old Rafael Nadal had recovered from an injury and defeated 23 year old Caspar Ruud in the French Open. Nadal’s 22nd Grand Slam win. Nadal has the best resources in the world to keep him in top form, but even the tennis great has dealt with physical set backs. And then there is the rest of us, who must forge on, without a world class team to keep us moving. Could AI, and a Tonus Tech wearable robotics suit help all of us remain active and healthy into our latest decades?
I’m betting that Marianna’s intelligence, energy and enthusiasm will help us avoid being forced to the sidelines, allowing us to keep moving and enjoying the activities that we love.
You can contact Marianna at firstname.lastname@example.org. https://www.tonus.tech/