Researchers from Northwestern Medicine and Rehabilitation Institute of Chicago (RIC) have made smartphone activity trackers smarter than ever; they have trained trackers to have the ability to detect the difference between fake and real activity.
Insurance companies and health care providers base rewards to individuals who stay active and healthy on the results from wearable activity trackers, among them smartphones. They also rely on these devices to monitor a patient’s activity and determine whether patients are following through on their physician’s recommendations to improve their physical health and/or the outcome of a treatment.
Current activity trackers can easily interpret fake activity, such as deliberately shaking the phone, as real activity like taking a brisk walk.
Northwestern researchers were able to train systems on data obtained through this fake, deceptive behavior, to increase the devices’ accuracy to 84 percent. The trained devices were able to identify this behavior and then generalize it for all individuals. Simply put, if a tracker can find out how one person “fakes” their activity data, it will be able to identify similar behavior that others participate in.
Northwestern’s study was published in the PLOS ONE journal in December 2015. As the senior author on the paper, research scientist at RIC and an associated professor in physical medicine and rehabilitation, Konrad Kording, notes, “Very few studies have tried to make activity tracking recognition robust against cheating.” He added, “This technology could have broad implications for companies that make activity trackers and insurance companies alike as they seek to more reliably record movement.”