Falls are dangerous among the elderly population and are a major health concern. Many investigators and some commercial products have used accelerometers for real-time fall detection. Furthermore, the combination of miniature 3-D accelerometers and gyroscopes has also been used for fall detection sensors, but the effects of the optimal placement position of these sensors have not been studied thoroughly. Besides, intelligent algorithms can be used to improve the accuracy of these wearable fall detection sensors considerably. Therefore, in this chapter, we will report two related fall detection studies that used slightly different sensor technologies and smart algorithms as follows: (1) the 1st study reports our custom-designed wired sensor (with both an accelerometer and two gyroscopes) placed as a single unit on three different positions along the thoracic vertebrae on subjects (i.e., T-4, T-7, and T-10). Results indicated that T-10 was not a good location for fall detection; however, both T-4 and T-7 were suitable, with the results for T-4 being slightly better. Using a simple rule-based multi-thresholds algorithm that utilizes the recorded resultant gravitational acceleration, angular change, angular velocity, and angular acceleration, we were able to successfully detect all 60 falls and differentiate between falls and activities of daily living (ADL) with no false positives on young volunteers (note: backward falls were not included in this sensor placement study as we were using a wired sensor). This naturally leads to our second study: (2) a real-time fall detection system using our improved wireless gait analysis sensor (WGAS) will be detailed. This WGAS is capable of differentiating falls vs. various Activities of Daily Living (ADL) and the Dynamic Gait Index (DGI) tests performed by young volunteers using a Back Propagation Artificial Neural Network (BP ANN) algorithm. The WGAS, which includes a tri-axial accelerometer, 2 gyroscopes, a wireless module and a Texas Instruments (TI) MSP430 microcontroller can be worn by the subjects at either T4 (at back) or on a belt-clip (in front at the center of the waist) for performing the various falls, ADL, and DGI tests. The raw gait data is wirelessly transmitted from the WGAS to a nearby PC for real-time fall classification, where six features were extracted for the BP ANN. Overall fall classification accuracies of 97.0% and 97.4% have been achieved for the data taken at the T4 and at the belt positions, respectively. The preliminary data demonstrates an overall sensitivity of 97.0% and overall specificity of 97.2% for this WGAS fall detection system, suggesting this can be a good candidate as a real-time low-cost and effective fall detection device for wireless acute-care and wireless assisted-living. We are now collecting more clinical data using the WGAS system on both patients and volunteers for gait analysis targeting for fall prevention as well.
|Title of host publication||Activities of Daily Living (ADL)|
|Subtitle of host publication||Cultural Differences, Impacts of Disease and Long-Term Health Effects|
|Publisher||Nova Science Publishers, Inc.|
|Number of pages||26|
|State||Published - Jul 1 2015|