Estimation of Passive Wrist Stiffness Across Two Coupled Degrees of Freedom Using a Kinematic Model of The Human Wrist

摘要

The evaluation of neurological disorders, such as Parkinson's Disease or stroke, relies on the examination of joint rigidity. Clinical rigidity can be quantitatively assessed by the estimation of the static passive component of the joint impedance, known as passive stiffness. Here we propose an experimental setup and a methodology for estimating the passive stiffness of the human wrist across two coupled Degrees of Freedom (DoFs): Flexo-Extension (FE) and Radio-Ulnar Deviation (RUD). A robotic arm passively moved the participant's wrist, whose movements and interaction forces/torques were recorded through infrared cameras and a 6-axis load cell, respectively. Conversely to most previous studies, that approximated wrist angular displacements and torques with the robot joint angles and torques, we considered a wrist kinematic model with two orthogonal axes (FE and RUD) intersecting at the wrist center of rotation. We tested our method on four healthy volunteers and employed multiple linear regression to estimate 2-by-2 stiffness matrix and represent its symmetric component in the form of a stiffness ellipse. The results were comparable with the prior literature. Specifically, the joint stiffness in the FE-RUD space was confirmed to be anisotropic, with the stiffness along RUD higher than along FE, and with the axis of maximum stiffness tilted counterclockwise of a few degrees with respect to the pure RUD axis.