Toward dynamic rehabilitation management: A novel smart product-service system development approach based on fine-tuned large vision model and Fuzzy-Dematel
Wenyu Yuan,Hua Zhao,Xiongjie Yang,Ting Han,Danni Chang
DOI: https://doi.org/10.1016/j.aei.2024.102616
IF: 8.8
2024-06-09
Advanced Engineering Informatics
Abstract:Nowadays, transformative technologies such as artificial intelligence, big data, and cloud computing are significantly influencing and reshaping the daily lives of individuals. Guided by the overarching concept of digital transformation, data-driven Smart Product-Service Systems (SPSS) have emerged, prompting scholars to investigate development approaches tailored to diverse data sources. However, the current approaches employed in the construction of SPSS exhibit limited capability in processing vast amounts of user-generated unstructured data. The relationship between big data intelligence and personalized services remains undisclosed. Moreover, the current focus of SPSS orientation predominantly addresses end consumers or manufacturers, with inadequate attention given to dynamic collaborative models that involve multiple stakeholders. These gaps are particularly conspicuous in complex industries such as rehabilitation management. To tackle these challenges, this study introduces a novel SPSS development approach that integrates a large vision model and the fuzzy-DEMATEL method. Specifically, a data-driven predictive assessment module was proposed, which constructs a medical image dataset and trains a rehabilitation predictive assessment model based on the transformer architecture. Secondly, personalized intervention services were generated, involving the representation of system elements, configuration, and optimization of service parameters. The fuzzy-DEMATEL method is mainly used for the initialization of service parameters. Then, interactive feedback is integrated into rehabilitation exercises for achieving continuous rehabilitation evaluation and service improvement. To validate the proposed approach, a FPRM-SPSS case was implemented, and it shows that the predictive assessment model achieved a high level of accuracy when applied to the clinical dataset constructed in this study, and the system was evaluated with high scores in user satisfaction.
engineering, multidisciplinary,computer science, artificial intelligence