Detection of bouncing during sit-to-stand transfers with sequential pressure images

As adults age or suffer from physical ailments, their level of functional strength and mobility can decrease over time. One indicator of a loss of muscle strength is the occurrence of bouncing during the sit-to-stand sequence while exiting a bed. This paper presents two algorithms designed for the detection of bouncing from pressure images and tests them on segments of healthy, simulated bouncing and post-stroke patients. The first algorithm relies on pressure measurements from the regions of contact by the hips and hands. This algorithm is found to be most accurate when the hands are both placed on the bed at a distance from the hips but decreases in accuracy when they are either placed too close, on top of the thighs or not used at all. To compensate for this, the second algorithm considers the full image as one region and measures the centroid along with the total pressure over time. This algorithm is successful at bounce detection regardless of the region definition, but can lose accuracy when the patient pushes off the bed with little forward lean. The intelligent fusion of these two algorithms within a mobility monitoring system can provide the detection of bouncing in a wide range of occupants within the smart home environment.