Although she wears no Superman cape, Dr. Dina Katabi has superhero powers. RF-Pose, a breakthrough new x-ray technology she invented, allows real-time imaging through solid objects. With RF-Pose, Katabi can literally see through walls.
A native of Damascus, Syria, Katabi has turned the science fiction dreams of her childhood into reality. “When I was a kid, I was very much fascinated by Star Wars,” she remembers. “And what fascinated me most is the notion of the Force, this energy that connects all people and all objects and allows you to feel people that you can’t even see.”
Holding these fantastic ideas in her heart, Katabi pursued a career in electrical engineering. There, she worked closely with ‘invisible’ forces like radio waves and electricity. In 2017, she received the Association of Computer Machinery Prize in Computing for her work in wireless signaling and wireless coding systems, and teaches at MIT’s Computer Science and Artificial Intelligence Laboratory.
Dina Katabi Deploys Cutting-Edge Technology
Katabi’s newest engineering inventions uses artificial intelligence to estimate the relative position of a human being. This AI can make its estimations even from the other side of physical objects – like walls. This breakthrough technology has numerous potential applications, including enabling caretakers monitor the safety of nursing home patients from a distance.
RF-Pose works by sending radio waves through walls and other obstacles. These waves reflect off the human’s neural system to display a virtual map of the human skeleton. In order to translate the interactions of radio waves and human limbs into observable images, Katabi and her team synced cameras with the wave-generating apparatus.
Eventually the neural network learned to connect the waves to the corresponding motions, even specific human movement. Although the visual display is more of a stick figure than detailed skeleton, the RF-Pose can estimate a person’s position with enough accuracy to tell when a patient may be in danger or in need of help.
What RF-Pose Can Do
Katabi sees more than one application for the RF-Pose. For one, doctors can use the technology to continually monitor patients with chronic diseases, so that they can better adjust prescription dosages and treat the condition in its early stages.
Beyond this, the RF-Pose offers opportunities to monitor the sick and elderly without invading their privacy. Staff could detect if an elderly patient was in an uncomfortable or dangerous position inside their private room. If a person falls in their own bathroom and is unable to call out for help, RF-Pose would be able to alert residence staff of an abnormal change in posture.
Outside of the field of healthcare, the RF-Pose has potential for aiding search-and-rescue teams. Additionally, it could someday offer the basis for a more refined virtual reality technology. In this case, players could simply move their bodies to perform corresponding digital actions.
Dina Katabi Connects with Her Roots
Despite her many successes an MIT professor Katabi still holds her origins close to heart. Though she has not lived in Syria since earning her bachelor’s degree from the University of Damascus in 1995, she finds inspiration in her Levantine roots.
“There’s an author who is very famous, Tawfiq al-Hakim,” she observes. “He is very good at abstraction, which is exactly what we do as engineers and mathematicians. We want to abstract something that is very complex and put it in simple terms.”
Like al-Hakim, the ability to condense abstract concepts into tangible actions and inventions has launched a successful career for Katabi, now the recipient of several awards, including the MacArthur Foundation “genius grant”, and has been referenced over 20,000 times in academic papers.
With Katabi’s contribution, neural networks can now detect the relative positions and changes of people behind walls. But that doesn’t mean Katabi is satisfied. She continues to work on improvements to the technology, in the hopes that RF-Pose will learn to recognize even fine motor motions in 3 dimensions.