The way that a person walks can contain important information about their health. In 2011, a University of Pittsburgh study found that walking speeds could predict life expectancy. Now, MIT researchers have found an easy way to monitor how fast people are walking at home, as The Verge reports.
The MIT Computer Science and Artificial Intelligence Laboratory (CSAIL)—which has previously debuted technology to help people learn languages and evaluate their selfies—is working on a wall-mounted sensor that can track people’s movements via radio signals. WiGait sends out low-radiation signals that reflect off a person’s body as they walk through the room, then calculates how fast that person is moving based on how the signals come back. It’s 95 to 99 percent accurate at measuring people’s walking speeds and stride lengths, according to MIT. The team behind the device claims it is more accurate than trackers like FitBit, which don't measure velocity or stride length, just step numbers. WiGait can recognize the movements of up to four different people in a room and works at distances of up to 40 feet, including through walls.
The technology could be used to study older people’s walking speeds at home to analyze their health without requiring them to wear an intrusive wristband or other wearable device. Walking speed can be an indicator of a person’s risks for future falls and hospitalizations, and WiGait could be used to monitor patients with a high risk of falling or cognitive decline, especially in assisted-living homes.
The CSAIL research [PDF] tested out WiGait in 14 different homes, and found that users tended to forget it was there after a few days, and preferred it to the idea of installing a camera-based monitoring device in their home. Most people don’t want to be recorded 24/7, but radio signals that only recognize movement might be a more acceptable way to collect data on people’s gait speeds at home, where, in the case of older people with limited mobility, they probably spend most of their time.
[h/t The Verge]