Dynamic Parameter Identification of a Human-Exoskeleton System With the Motor Torque Data

The objective of this paper is to identify dynamic parameters of the human-exoskeleton system (HES) with the motor torque data considering modeling the passive elastic joint torque of the user. Based on a lower limb exoskeleton prototype we recently developed, the dynamics of the joint actuators and two degrees of freedom link-based swing leg are firstly modeled. The passive elasticity of the human joint is modeled by double-exponential equations. Systematic experiments are then conducted, during which the motor current is measured to compute the motor torque for identification. Finally, the dynamic parameters of the joint actuators, the exoskeleton leg and the human-exoskeleton leg are successively identified. The identification is implemented through the off-line inverse dynamic model simulation with the measured motor torque data. A consistent tendency is found between the simulated motor torque after identification and the measured one. This demonstrates the validity of the identification method proposed. This paper provides a systemic method of identifying the HES with the motor torque data and enriches the experimental proof that the double-exponential model is feasible to describe the passive elastic joint torque.