Validity of four gait models to estimate walked distance from vertical COG acceleration

Pedometers are basically step counters usually used to estimate the distance walked by a pedestrian. Although their precision to compute the number of steps is quite accurate (about 1%), their feasibility to estimate the walked distance is very poor, as they do not consider the intrinsic variability of human gait. Reported results show values of 10% of precision in optimal conditions, increasing to 50% when conditions differ. Electronic accelerometer-based pedometers base their functioning on a basic processing of the vertical acceleration of the waist. Recently, different approaches have been proposed to relate such signals to the step length. This can lead to an improvement of the performance of this kind of device for estimating the walked distance. In this article, we analyze four gait models applied to the vertical accelerations of the body's center of gravity, three biomechanical and one empirical. We compare their precision and accuracy. Results support the superior performance of three of them over an ideal pedometer. We also analyze their feasibility to be implemented in pedometer-like devices.