Measuring a person’s walking speed using a wearable device is actually a whole lot more difficult than you might think. For example, devices such as FitBits base their estimate of how fast a person is walking solely on the number of steps taken, resulting in numbers that aren’t always accurate.
A new research project carried out at the Massachusetts Institute of Technology (MIT) Computer Science and Artificial Intelligence Laboratory (CSAIL) claims to have cracked this challenge, however — with a system that can measure the walking speed of multiple people with 95 to 99 percent accuracy. Most impressive of all? It doesn’t require any wearable tech whatsoever.
Instead, it involves a wall-mounted sensor called WiGait that can be placed in a person’s home. By analyzing the wireless signals reflected off a person’s body, the team is able to make its accurate predictions. It is also able to ascertain a person’s stride length with 85 to 99 percent accuracy.
But while this certainly an impressive computer science demo, it’s what the team wants to do with this data that’s really exciting.
“This builds on our earlier work measuring breathing and heart rate by analyzing the surrounding wireless signals, without any wearables,” lead author Chen-Yu Hsu told Digital Trends. “We picture a connected world where we can equip our homes to use ambient wireless signals to monitor our health, from tracking metrics related to chronic diseases to alerting us about health emergencies. WiGait’s ability to monitor walking speed, stride length, and changes in mobility habits is a big leap in that direction.
In this scenario, being able to measure walking speed and stride length could provide useful health insights. This is something already being backed up in independent studies. For instance, features like cognitive decline and possible cardiac disease could be sensed by looking at the speed at which a person moves. In the future, the team also hopes to train it on people suffering from walking impairments such as Parkinson’s disease, which could have its progression tracked by a person’s stride length, since the disease is characterized by small shuffling steps.
“Many of us worry about the health of our parents and grandparents,” Hsu continued. “We see this as filling a need for in-home non-invasive health monitoring that can give us peace of mind about our loved ones’ health. We believe that having a smart device like this that analyzes the wireless signals can address some of the key issues we face in healthcare.”
