Accurate estimation of finger joint stiffness is important in assessing the hand condition of stroke patients and developing effective rehabilitation plans. Recent technological advances have enabled the efficient performance of hand therapy and assessment by estimating joint stiffness using soft actuators. While joint modular soft actuators have enabled cost-effective and personalized stiffness estimation, existing approaches face limitations. A corrective approach based on an analytical model suffers from actuator–finger and inter-actuator interactions, particularly in multi-joint systems. In contrast, a data-driven approach struggles with generalization due to limited availability of labeled data. In this study, we proposed a method for energy conservation-based online tuning of the analytical model using an artificial neural network (ANN) to address these challenges. By analyzing each term in the analytical model, we identified causes of estimation error and introduced correction parameters that satisfy energy balance within the actuator–finger complex. The ANN enhances the analytical model’s adaptability to measurement data, thereby improving estimation accuracy. The results show that our method outperforms the conventional corrective approach and exhibits better generalization potential than the purely data-driven approach. In addition, the method also proved effective in estimating stiffness in human subjects, where errors tend to be larger than in prototype experiments. This study is an essential step toward the realization of personalized rehabilitation.

The Myoshirt, an active exosuit, provides gravity compensation for the shoulders. This study evaluated the impact of the Myoshirt on range of motion (ROM), endurance, and activities of daily living (ADLs) performance through tests involving nine participants with varying levels of arm impairments and diverse pathologies. Optical motion capture was used to quantify ROM of the shoulder and elbow joints during isolated movements and functional tasks. Endurance was quantified through a timed isometric shoulder flexion task, and a battery of ADL tasks was used to measure the perceived support of the exosuit, along with changes in movement quality. Feedback and usability insights were gathered with surveys. The Myoshirt did not significantly improve ROM during isolated movements (shoulder flexion, shoulder abduction, and elbow flexion/extension), but during the reaching phase of a functional drinking task elbow extension increased significantly by 13.5% (t = 7.52, p = .002). Participants could also keep their arms elevated 78.7% longer (t = 1.942, p = .047). Patients also reported less perceived difficulty with ADLs while using the device, and a therapist reported improved execution quality. Participants who self-reported severe impairment levels tended to derive greater benefits compared to those with milder impairments. These findings highlight the potential of the Myoshirt as an assistive device, particularly for individuals with severe impairments, while emphasizing the need for further refinement.