An algorithm to decompose ground reaction forces and moments from a single force platform in walking gait

In walking experimental conditions, subjects are sometimes unable to perform two steps on two different forceplates. This leads the authors to develop methods for discerning right and left ground reaction data while they are summed during the double support in walking. The aim of this study is to propose an adaptive transition function that considers the walking speed and ground reaction forces (GRF). A transition function is used to estimate left and right side GRF signals in double support. It includes a shape coefficient adjusted using single support GRF parameters. This shape coefficient is optimized by a non-linear least-square curve-fitting procedure to match the estimated signals with real GRF. A multiple regression is then performed to identify GRF parameters of major importance selected to compute the right and left GRF of the double support. Relative RMSE (RMSER), maximum GRF differences normalized to body mass and differences of center of pressure (CoP) are computed between real and decomposed signals. During double support, RMSER are 6%, 18%, 3.8%, 4.3%, 3%, and 12.3% for anterior force, lateral force, vertical force, frontal moment, sagittal moment and transverse moment, respectively. Maximum GRF differences normalized to body mass are lower than 1N/kg and mean CoP difference is 0.0135 m, when comparing real to decomposed signals during double support. This work shows the accuracy of an adaptive transition function to decompose GRF and moment of right and left sides. This method is especially useful to accurately discern right and left GRF data in single force platform configurations.