Sensors' Ground Reaction Force behavior for both Normal and Parkinson subjects - A qualitative study

Characterization of normal and abnormal Gait has been a major research field for decades, whether in fall prevention, sports biomechanics or even disease indication. In this paper, we assess time domain statistical properties of the Vertical Ground Reaction Force (VGRF) during moderate-pace walking, aiming eventually to create a reliable mathematical model of VGRF for normal and abnormal cases. For that endeavor, first order statistical analysis was performed upon signal segmentation in order to determine the degree of stationarity and base the model upon it. Furthermore, we performed curve fitting of the VGRF time series between present and past values, which led us to model the waveform with linear regression via Autoregressive Model for both Normal Walking Signals and Parkinson diseased patients' walking signals. However this is done only for one chosen sensor. However, it would be crucial to take the advantage of the array of sensors. Evaluating the cross-covariance between multi-sensor data of a given subject at different time lags capture the most important information. The seasonality in the values give a quite important indications of the behavior of data. The objective behind this analysis is to recommend a preliminary basis to create reliable mathematical model of normal walking signals versus pathological walking signals, that we will emphasize in a complementary work, in the simplest way available and creating fall prevention indicators for old patients.