Hand function impairment is a common consequence of neurological diseases. Conventional rehabilitation approaches often suffer from slow progress and large inter-individual variability. Hand exoskeletons have become an important rehabilitation technology, but they still have shortcomings in dexterity, intention recognition, and natural interaction. This article focuses on the development of hand exoskeletons from being unskilled to dexterous, systematically reviewing their evolution in mechanical structure, driving methods, sensor integration, and control algorithms. Through a comparison of technology and clinical applications, it was summarized that it has transitioned from rigid structures and single-drive mechanisms to flexible materials, multimodal sensing, and intelligent control. And generalized the breakthrough achievements in key areas such as tendon-driven mechanisms, soft structures, biological signal analysis, and human-computer interaction. Research has confirmed that flexible actuation, multimodal sensing, and intention-based adaptive control strategies can significantly enhance the naturalness, safety, and multi-degree-of-freedom capabilities, thereby encouraging patients to engage in active training and accelerating the process of neural plasticity reconstruction. The conclusion points out that dexterity is the core factor driving the lightweight, flexible, intelligent design of hand exoskeletons and precise clinical rehabilitation. In the future, there is a need to strengthen the integration of medicine and engineering, as well as standardized assessment.
Zhipeng Jiang (Mon,) studied this question.