New attempt of proposing the pedometer algorithm in the elderly

A pedometer is popular as a quantitative evaluator of walking. However it may not measure the number of steps for elderly. In this study, we attempted to propose a step count algorithm based on a filter bank and a threshold processing to improve the accuracy of an accelerometer-type step counter for elderly. The waist acceleration signals during a self-paced walking were recorded form 74 participants of the gait training at the commuting rehabilitation services center using triaxial accelerometer. A conventional accelerometer-type pedometer was also attached close to the accelerometer. They walked approximately 20 m. As a result of the algorithm verification using 74 elderly data sets, 57 subjects could be estimated with less than 10% error. We suggest that our step count algorithm is suitable for estimating the number of steps for elderly.