Gait segmentation using continuous wavelet transform for extracting validated gait events from accelerometer signals

In our aging society, gait disturbance becomes a major concern as it inevitably leads to limitations in mobility and to an increased risk of falls. Spontaneous walking speed normally decreases by about 1% per year from age 60 onward [1]. A population-based study has shown a 35% prevalence of gait disorders among persons over age 70 [2], and some abnormal gait features can be predictive of a progression to dementia of Alzheimer type, e.g., [3], [4]. Problems of balance and gait are associated with immobility and falls, which markedly impair the quality of life, e.g., [5]. Thus, it is important to develop quantitative methods aimed at monitoring gait disturbances in natural environments, and at the critical analysis and development of new therapeutic strategies. Accelerometer-based methods have become popular to deal with such applications using accelerometers with small size and low power consumption, as well as algorithms that accurately extract relevant gait events and gait phases, e.g., [6], [7].