A method of walking parameters estimation via 3-axis accelerometer

Walking parameters (walking speed, stride period, walking distance) are very important for the stroke rehabilitation assessment and guide because the walking parameters of stroke patients in contrast to normal person change a lot and they are great quantitative indexes of walking capacity. In order to detect the stroke patients´ walking parameters, a method of real-time walking parameters estimation with a single accelerometer is proposed in this study. The zero-approximation step detection algorithm is employed to segment the continuous acceleration in terms of foot, up and down. Accelerations at different walking speeds are recorded to investigate the method´s viability. Accelerations on different axis are employed to obtain accurate walking parameters. The average errors of walking speed and distance of tests estimated with the sum of the acceleration output are 4.7% and 5.8% respectively. The experimental results indicate that the proposed method is robust and exact.