Fall detection by smart phone in older adults

MSR145 data loggers in use for a research of the University of Roehampton, UK

Authors: Raymond Y. W . Lee (PhD) and Alison J. Carlisle (MSc), University of Roehampton, UK

Researchers at the University of Roehampton Musculoskeletal research group used MSR145 data loggers with tri-axial accelerometers to record signals during a fall of the human body.

Falling among older adults can cause major distress; the consequences may include fracture and the inability to move and isolated individuals may experience a long lie after a fall when they are unable to call for help. To alleviate this Professor Raymond Lee and colleagues developed and tested a smart phone application. For instance, the acceleration signals of the MSR data logger can be acquired by a smart phone and then compared with pre-defined thresholds. The application will send out a distress signal to call for help when a fall has been detected. The ‘Fall Detection Application’ can be adjusted to suit an individual’s profile by adjusting individual acceleration thresholds, after which it could be used to detect a fall. If this was to happen a loud beeping and warning text could be sent automatically to summon help. The diagram below (Figure 1) illustrates the concept:

Figure 1: A block diagram illustrating the fall detection algorithm developed for the mobile phone.

It is also possible to detect falls using the accelerometer of the phone, but this requires the user to have the phone attached to the waist of the body at all times which incurs the obvious problem of the user forgetting to keep the phone on the waist. False alarms might occur if the phone is dropped compromising specificity and sensitivity. Therefore using an external accelerometer, like the MSR145 datalogger, may be a better option. In fact, the researchers found the acceleration signals from the MSR145 had a higher accuracy than those from the embedded accelerometer inside the phone.

Full details of the research study can be found in the scientific publication in the journal Age and ageing.