Prompt fall detection in any living space, including bathrooms, can save lives.
By George Shaker
magine a future where your home could monitor your health and daily activities without cameras or wearable devices.
There was a time this type of technology or idea only existed in science fiction, but researchers at the University of Waterloo and the Schlegel-UW Research Institute for Aging (RIA) have developed a new system that can do just that using wireless signals.
The team of researchers, led by George Shaker, PhD, adjunct associate professor, University of Waterloo created a state-of-the-art demonstration facility (MIRADA) at the RIA that allows researchers to test technology-driven solutions designed to help older adults take control over their health and well-being.
The Monitoring, Intervention, and Response for Aging Demo Apartment (MIRADA) is a space where Schlegel Research Chairs and collaborators can study and demonstrate groundbreaking sensing technologies, advanced monitoring systems, and timely interventions to help address the unique challenges faced by aging populations.
The research team recently demonstrated an in-home monitoring system that uses radar sensors and artificial intelligence to detect a person’s location and movements within a home without the use of cameras or wearable devices. The sensor emits low-power radio waves – less than those emitted by a WiFi system – that reflect off a person’s body. When combined with machine learning algorithms in the cloud, this data can identify specific activities like walking, sleeping, eating, and even falling.
According to Shaker, this technology has exciting potential for healthcare, especially for older adults.
“Continuously monitoring health and activities at home could allow for early detection of changes that indicate declining health. This would enable proactive interventions that could prolong independent living,” he explains.
Prompt fall detection in any living space, including bathrooms, can save lives. The technology can also monitor bathroom visit frequency and duration, which provides insight into hydration status and potential onset of medical conditions such as, urinary tract infections.
Beyond falls and bathroom monitoring, gait changes often precede cognitive and physical decline. The radar-based system can unobtrusively measure walking speed and other gait metrics daily rather than periodically at clinic visits. Doctors can use the system to monitor subtle gait changes, triggering a closer evaluation when needed.
The system’s activity recognition capability provides insight into functional health. Decreased time upright and walking may reflect emerging mobility limitations or depression. From a safety perspective, the technology holds promise for monitoring high-risk individuals without
intrusive cameras.
Beyond health applications, the technology could also help smart home devices better support people by adjusting to their activities and needs.
Evaluating the system
To develop and evaluate the system, Shaker’s team installed radar units in MIRADA – which resembles a typical multi-room living space. Participants performed daily living activities like walking, washing dishes, vacuuming, and sitting, which generated over 300,000 radar data samples. Artificial Intelligence algorithms extracted informative features from the radar signals to classify the different activities. The researchers tested several types of artificial neural networks on the radar data before finding a candidate network that achieved excellent
performance.
When tested on new subjects, the system accurately identified activities like sitting, washing dishes, and walking around 87 per cent of the time.
“A key advantage of our machine learning approach is the ability to learn what motions characterize particular activities across individuals without needing complex rules programmed by hand,” he explains.
While further refinements are needed before real-world use, the system represents an important step towards accessible in-home health monitoring. Shaker believes the technology could be commercialized using inexpensive radar sensors like those in smartphones paired with cloud-based AI services.
The biggest barrier may be user acceptance of radar monitoring despite its unobtrusive nature. The research team is exploring how to communicate the technology’s privacy-preserving benefits and healthcare value proposition to help drive adoption. “Wireless sensing technologies stand to transform home-based healthcare,” believes the research team. “However, we need to ensure these innovative technologies are implemented in a human-centered way that promotes comfort and autonomy.”
With its healthcare potential and ability to assess gait, detect falls, and monitor activities, this radar monitoring system offers an exciting glimpse into how wireless sensing and artificial intelligence could support healthier independent living. While further research is still required, the future looks bright for in-home technologies that enhance care options without jeopardizing privacy.
George Shaker, PhD is an adjunct associate professor, University of Waterloo; Research scientist, Schlegel-UW Research Institute for Aging.
