Finger Joint Stiffness Estimation with Joint Modular Soft Actuators for Hand Telerehabilitation

Abstract

In a telerehabilitation environment, it is difficult for a therapist to understand the condition of a patient’s finger joints because of the lack of direct assessment. In particular, not enabling the provision of spasticity evaluation significantly reduces the optimal performance of telerehabilitation. In a previous study, it has been proposed that finger stiffness could be estimated using an analytical model of a whole-finger soft actuator. However, because the whole-finger soft actuators require high air pressure for high bending performance and are costly to customize for each patient, using joint modular soft actuators for telerehabilitation turns to be a necessity, though stiffness estimation with joint modular soft actuators has not been studied yet. Another problem is caused by using a marker-based joint angle measurement, which requires the markers to be attached to the exact positions, and limits its application in telerehabilitation. In this study, we proposed a procedure of finger joint stiffness estimation that combines information acquired from a joint modular soft actuator and a marker-less hand joint position acquisition device. Correction parameters were added to the previous analytical model for -the bending analysis of a joint assisted using a joint modular soft actuator. Moreover, a multi-variate regression model was implemented for correcting joint angles obtained from the hand joint position acquisition device. As a result, a reasonable accuracy of stiffness estimation was achieved for rehabilitation with the joint modular soft actuators, which suggests the possibility of using the proposed method to evaluate the finger spasticity in a telerehabilitation environment. This is a big step forward towards optimal hand telerehabilitation.

Deriving Design Rules for Personalization of Soft Rehabilitation Gloves

“…2). The difference in lengths chosen as the test cases was ± 5 and 10 mm based on 20 mm, which was designed as the standard size in previous studies [3], [22]. This length can also be modified according to the design rules described in SectionII-C.…”

Addressing data imbalance in Sim2Real: ImbalSim2Real scheme and its application in finger joint stiffness self-sensing for soft robot-assisted rehabilitation

“…The finger joint stiffness data in the simulation were 0.11, 0.15, 0.2, 0.5, 0.58, 0.7, 1.03, 1.19, 1.4, 1.70, and 2.12 Nmm/° and in real world, were 0.11, 0.58, 1.03, 1.19, and 2.11 Nmm/°. These values were selected based on the range of finger joint stiffness observed in individuals with spasticity and in healthy individuals ( Matsunaga et al, 2023 ). Among them, 0.11, 1.19, and 2.12 Nmm/° were selected as the training dataset, and 0.58, 1.53, and 1.03 Nmm/° were selected as the validation dataset.…”

“…In particular, Heung et al analyzed the relationship between finger joint stiffness, soft actuator’s angle, and air pressure, and hence developed an accurate analytic angle-pressure-finger joint stiffness model ( Heung et al, 2020 ). However, this approach is based on the chamber structure of the soft actuator, which cannot be used with a model-unknown soft actuator because of its model dependency ( Matsunaga et al, 2023 ). In spite of the lack of the literature, it is not difficult to conceive the possibility of estimating finger joint stiffness by training a neural network-based finger joint stiffness self-sensing scheme with the air pressure and angle of a soft actuator as inputs.…”

Development of a Dual Function Joint Modular Soft Actuator and its Evaluation Using a Novel Dummy Finger Joint-Soft Actuator Complex Model

“…PREPRINT VERSION. ACCEPTED FEBRUARY, 2024 without taking into consideration the interaction with a dummy joint [12]. One challenge when simulating the interaction between fingers and soft actuators is that it involves dealing with multiple soft and rigid bodies, which can be complex as it requires the simulation of friction between contact surfaces.…”