Predicting the dynamics of the human gait single support phase

Abstract

Simulations of human body motion involve different fields of study such as robotics, from nano to macroscale, and biomechanics. The simulation of human motion is a challenging problem from the physical and computational perspectives and several models have been proposed in literature. Using a methodology named predictive dynamics, this work aims to propose a 2D spatial model of human walking during the single stance phase (SSP), using a flat foot adaptation. The model is based on inverse dynamics in which the angular displacements are interpolated by 5th degree B-splines. The equation of motion is developed by the recursive Lagrangian formulation due to its computational efficiency. An optimization problem is set up to obtain the control points of the B-splines using as the objective function the dynamic effort, which is a limited power supply. The constraints imposed on the motion are time-dependent, such as the torque/angle limits, and the dynamic stability criterion defined by the zero moment point (ZMP), or time-independent, such the equation of motion and step length. All the steps are explained in detail to find the dynamics of the body using the smallest number of laboratory results. The solution from the model is favorably compared to Winter’s data, in particular the ground reaction forces, which is important in many engineering applications.