Anticipating accidents: Petr Dobiáš’s vision for smarter elderly care

An aging population is one of the most pressing challenges facing modern society. In the second installment of Radio Prague International’s new edition of Science without Borders, this time brought to you from France, Petr Dobiáš explains how engineering, computer science, and artificial intelligence intersect in the development of a system that can anticipate emergencies, not just react to them, without sacrificing human dignity or privacy.

When would you say that technology and research first started capturing your attention?

“I think it was since my early childhood, because when I was young, I observed different phenomena, nature, technology... I asked a lot of questions to my parents, and they always answered me with all the details, so I think it's since my early childhood.”

You are currently working at the ETIS Lab. Could you describe to us what kind of venture it is and what its missions are?

“My laboratory, ETIS Lab, is a research lab, that is public. There are three main sources of funding. The first one is CNRS, so the French Research Institute, the second one is my university, CY Cergy Paris Université, and the third one is the electronic engineering school ENSEA. So these are the three main fundings. This laboratory was created in 1980, and currently there are 170 members, and approximately 60% of them are PhD students.”

Now, for your research project, particularly, what are the main thoughts? What are the main principles behind it?

“So my research is within the project called SmartGait. And the aim of this project is to develop an intelligent system able to anticipate the falls of elderly people. What we would like to do is to use our system in the apartments of elderly people, or in the nursing home, in order to avoid these people needing to be in the hospital for the rest of their lives.”

Could you tell us more about the global issue that you're trying to solve with this and where maybe the current solutions fall a bit short?

“Project SmartGait tries to help solve the global issue of the aging of the population. In France, you have unfortunately more than 10,000 deaths of elderly people per year, and more than 136,000 hospitalizations per year of elderly people. So in the near future, with the population aging, it will be a severe problem to accommodate and help all these people who need to be treated in the hospital or at the nursing homes. Currently, there are two main solutions that can be used. The first one is to use sensors, such as watches or bracelets. The main problem is that there are a lot of false alarms, up to 90%. Another solution is to use a visual camera. But when using the visual camera, you do not respect the privacy of the monitored subject.”

So, could you tell us more about that solution that you are proposing and what information, what indicators you're working with that are invisible to the naked eye?

“Our project is based on a spatial imagery radar. We use the radar in order to protect the privacy of the monitored subject, so that nobody can recognize a person. Our solution is able to detect a fall, but this is the case that we would like to avoid. We aim to anticipate the falls by seeing in the radar data where there are indicators that show us that there is something deteriorating. And the main idea of our solution is basically that this radar, using a transmitter antenna, sends an electromagnetic wave. This electromagnetic wave is reflected by the subject, and these reflected waves are captured by receiver antennas. Then, it is rather heavy data processing to obtain the images that we call a spectrogram. It means speed as a function of time. These images are characteristic of each human activity. So it means that you have a special image for walking, you have a special image for sitting, for picking an object, and so on. And using these spectrograms, we can then foresee, recognize, and detect different human activities using the algorithm of artificial intelligence.”

Watch the video for the full interview

This interview was produced within a joint project by Radio Prague International, the Czech Centre in Paris, and the Czexpats in Science initiative. The Science Without Borders series with young successful Czech scientists working in various fields in France was recorded in Paris. It is a follow-up to the first series of ten interviews from Vienna