Key points are not available for this paper at this time.
During stroke rehabilitation, the recovery of upper-limb motor function is a primary therapeutic objective. Accurate rehabilitation assessment is essential for the scientific design and implementation of individualized training programs. However, current clinical practice predominantly relies on scale-based manual evaluations, which depend heavily on clinicians’ subjective judgment and may compromise objectivity and accuracy. To overcome this limitation, this study develops a novel horizontal upper-limb rehabilitation robot for quantifying patient movement data and proposes an upper-limb rehabilitation assessment model based on semi-quantitative information. The model integrates quantitative data collected by the robot with expert knowledge, employing a BRB algorithm to dynamically and objectively evaluate patients’ rehabilitation status. To validate the accuracy and effectiveness of the proposed method, BP and SVM neural networks are adopted for comparative models. Comparative analysis demonstrates that the semi-quantitative information-based BRB model achieves significantly higher accuracy than the alternative models. This approach not only enables rapid and precise monitoring of patients’ rehabilitation progress but also provides a robust foundation for the scientific formulation of individualized rehabilitation training plans.
Jiansheng et al. (Thu,) studied this question.