A Non-Wearable Approach to Fall Detection

Falls are a serious threat for senior citizens. In fact, according to the Centers for Disease Control and Prevention, one in every three people 65 and older falls each year in the United States. Two million U.S. seniors are treated in emergency departments for fall-related injuries annually.

To date, the main way falls have been detected is by fitting seniors with wearable radio/telephone-connected alarms. When a fall occurs, the senior presses an alarm button on the wearable to alert the monitoring station that they need help. Of course, should the senior be unconscious or too injured to press the button, this device is of no use to them—and no help will be forthcoming.

Now there are wearables that detect falls on their own accord and send calls for help without the intervention of the wearer. Unfortunately these devices still require seniors to actively wear them and ensure the devices are kept charged and working.

To address these deficiencies while collecting invaluable information on fall detection, prediction and prevention, researchers at the University of Missouri (MU) Center for Eldercare and Rehabilitation Technology have been experimenting on a range of non-wearable fall risk assessment (FRA) systems.

Initially the MU investigators installed FRA systems in 10 senior apartments at TigerPlace, a retirement community in the university town of Columbia. Conducted during 2011–12, this study saw each

TigerPlace apartment equipped with a Microsoft Kinect “depth camera” (originally developed for the Xbox gaming system) capable of detecting objects’ position in three-dimensional space. The apartments were also fitted with pulse-Doppler range-control radar (developed by General Electric) and two Web cameras.

After this initial study was done, the researchers realized they could capture the FRA data they needed by using the Kinect depth cameras alone. Since then the team has installed Kinects in a total of 150 senior apartments at TigerPlace and other senior housing.

“Our FRA systems provide continuous in-home monitoring of seniors in an unobtrusive, environmentally mounted system that automatically detects when falls have occurred—or when the risk of falling is increasing,” explains Marjorie Skubic, PhD, director of the Center for Eldercare and Rehabilitation Technology. Skubic is one of the study’s researchers and a professor in MU’s Electrical and Computer Engineering Department. “In this latter capacity,” she says, “the FRA system can alert family members and healthcare providers that the senior being monitored is showing signs of gait changes that could precipitate falls.”

The various measurements collected and predictions generated by the MU study have been validated by a number of accepted testing standards. They include the GAITRite electronic walkway gait measurement system.

Worth noting: The images captured by the Kinect depth cameras are converted to 3D silhouettes of the seniors being monitored. This protects them from feeling as if they are under surveillance.

The collected FRA data is analyzed using predictive algorithms that were tested and validated in the lab before the selected seniors’ apartments were outfitted. The resulting data not only detects falls but alerts researchers to indicators—such as changes in gait speed, stride time and stride length—that the senior might be becoming more fall-prone.

So far the results of the MU study have been quite encouraging. “We have been able to show and validate that it is possible to monitor seniors for falls without using wearable devices, and to also use ongoing monitoring to detect when they may become more prone to having falls due to declining physical health and fitness,” says Skubic. “We have also been able to see how actions such as transferring from certain models of wheelchairs can increase the risk of falls in some cases—and make preventative equipment changes as a result.”

At present the MU Eldertech research team continues its work at various senior residences. They are seeking to expand the testing of nonwearable FRA systems in other senior settings as well.

For healthcare providers who work with the elderly, the success of this nonwearable approach to fall management bears careful study, especially as it offers a real alternative to the current device-centric model. Find more details at https://eldertech.missouri.edu.

James Careless is a freelance writer with extensive experience covering computer technologies.