Fall Detection: Suggestions For Improvement

Alex Flitton
Oct 2, 2018 8:43:33 AM

Fall Detection is one of the most loved and hated features in PERS devices, to date. It’s overstated promise and underwhelming results make a lot of users, dealers and monitoring centers to pull their hair because of the number of false alarms it produces. However, that does not mean that fall detection is a lost cause. In fact, the technology has improved drastically over the years and continues to provide more meaningful results for subscribers. It only seems to appear in alternative methods, as opposed to the typical chest-level pendant.

Current Fall Detection Tech

The predominant version of fall detection technology is housed within a pendant that is worn around the necks of subscribers. According to several research papers from the last five years, the chest is the most reliable place on the body for fall detection. Because of its generally stable and slow-moving nature, the torso provides the most reliable information about the movement of the body.

Hands and wrists are poor areas for movement detection because of the frequent swinging and weight-bearing nature. Hands frequently make forceful connections with chairs, beds, countertops and automobiles. However, the torso is more stable in nature, so when a forceful connection is made with a surface, it is more probable to be with a floor, indicating a fall.

Gyroscopes and impact detection technology are two of the ways a device cross-references data to determine the nature of a subscriber’s movement. Therefore, if a subscriber makes an impact with a countertop, but the device doesn’t assume a horizontal position afterwards, it shouldn’t trigger an alarm. That doesn’t seem to be the case, however. In fact, you would be hard pressed to find a senior citizen who can finish doing the dishes without accidentally triggering their PERS device.

Alternative Fall Detection Methods

Companies like Kytera recognize a need to implement alternative methods for in-home fall detection technology. Their machine learning technology uses subscribers’ day-to-day movements to create a baseline for normal behavior. Therefore, when a subscriber behaves outside of the norm, the technology is prepared to recognize it. Their technology leverages three distinct methods for emergency detection:

  • An emergency button located on a wristband
  • Contextual activity tracking using sensors locating throughout a home
  • Routine-based distress detection

Although these state of the art technologies help effectively monitor individuals within a home, it doesn’t completely solve the typical problems associated with fall detection outside of the home.

Multiple-Parameter Triggers

No matter the direction that fall detection technology heads, it needs to improve. Still today, most signals that enter the central station are caused by false alarms. This creates a great amount of overhead and results in greater costs and unwelcomed legislative changes. Therefore, to improve the industry, grow profits and produce happier customers, fall detection needs a drastic overhaul.

You May Also Like

These Stories on PERS

Subscribe by Email

No Comments Yet

Let us know what you think