Optimized Gait Analysis Leveraging Wavelet Transform Coefficients from Body Acceleration

摘要

In this paper we present a new approach showing the feasibility of estimating kinematic features from the activity of a person wearing a single accelerometer on the waist. In particular, we show the correlation between the user's speed and the wavelet transform coefficients from the waist acceleration. This approach enables the minimization of hardware requirements for gait analysis and achieves results with errors below 10%, comparable to the most sophisticated systems available in medical environments.