Modal Identification on Walking Leg Using EKF Method

Abstract

The identification on the mechanical characteristic of walking leg has faced challenge due to the inconvenience installation on sensor with leg and expensive cost. This paper proposes a convenience method on identifying the leg parameters including its stiffness and damping ratio under walking state by Extend Kalman Filter (EKF) method, which only need to deal with the measured ground reaction force. Firstly, a mechanical oscillate system is used to describe the walking leg, and its corresponding dynamic governing equation is established. Then, a state-space equation is applied the governing equation and it is further inserted on the EKF method so as to set up a system update equation and measurement update equation. In addition, a solving algorithm is designed to acquire the instantaneous stiffness and damping ratio of the walking leg. Finally, a sample comprised of 760 curves of walking ground reaction force from 36 participants is employed to the identification method and the practice results demonstrate that it can efficiently discern the stiffness and damping ratio of walking leg. This study provides a novel perspective for seeking the dynamic parameters of walking leg and efficiently reduces experimental expenditure than those direct measurement of sensors installed on the surface of leg.