TY - GEN
T1 - Impedance Control of a 2-DOF Spherical 5-Bar Exoskeleton for Physical Human-Robot Interaction During Rehabilitation and Assessment
AU - Wolbrecht, E.
AU - Ketkar, V.
AU - Perry, J. C.
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - This paper presents a novel impedance controller for THINGER (THumb INdividuating Grasp Exercise Robot), a 2-degree-of-freedom (DOF) spherical 5-bar exoskeleton designed to augment FINGER (Finger INdividuating Grasp Exercise Robot). Many rehabilitation and assessment tasks, for which THINGER is designed, are improved by rendering near-zero impedance during physical human-robot interaction (pHRI). To achieve this goal, the presented impedance controller includes several novel features. First, a reference trajectory is omitted, allowing free movements. Second, force-feedback gains are reduced near actuator limits and a saturation function limits the maximum commanded force; both allow more responsive (higher) force-feedback gains within the workspace and mitigate transient oscillations caused by external disturbances. Finally, manipulability-based directional force-feedback gains help improve rendered impedance isotropy. Validation experiments included free exploration of the workspace, following a prescribed circular thumb motion, and intentional exposure to external disturbances. The experimental results show that the presented impedance controller significantly reduces impedance to subject-initiated motion and accurately renders the desired isotropic low-impedance environment.
AB - This paper presents a novel impedance controller for THINGER (THumb INdividuating Grasp Exercise Robot), a 2-degree-of-freedom (DOF) spherical 5-bar exoskeleton designed to augment FINGER (Finger INdividuating Grasp Exercise Robot). Many rehabilitation and assessment tasks, for which THINGER is designed, are improved by rendering near-zero impedance during physical human-robot interaction (pHRI). To achieve this goal, the presented impedance controller includes several novel features. First, a reference trajectory is omitted, allowing free movements. Second, force-feedback gains are reduced near actuator limits and a saturation function limits the maximum commanded force; both allow more responsive (higher) force-feedback gains within the workspace and mitigate transient oscillations caused by external disturbances. Finally, manipulability-based directional force-feedback gains help improve rendered impedance isotropy. Validation experiments included free exploration of the workspace, following a prescribed circular thumb motion, and intentional exposure to external disturbances. The experimental results show that the presented impedance controller significantly reduces impedance to subject-initiated motion and accurately renders the desired isotropic low-impedance environment.
UR - https://www.scopus.com/pages/publications/85176404998
U2 - 10.1109/ICORR58425.2023.10304762
DO - 10.1109/ICORR58425.2023.10304762
M3 - Conference contribution
C2 - 37941197
AN - SCOPUS:85176404998
T3 - IEEE International Conference on Rehabilitation Robotics
BT - 2023 International Conference on Rehabilitation Robotics, ICORR 2023
PB - IEEE Computer Society
T2 - 2023 International Conference on Rehabilitation Robotics, ICORR 2023
Y2 - 24 September 2023 through 28 September 2023
ER -