Scholarlink

<p>Improving upper limb motor deficits in severely disabled patients is a key therapeutic goal in rehabilitation. Recent clinical studies have shown that upper limb exoskeleton robots can facilitate repetitive training for stroke patients, aiding in regaining movement control and improving rehabilitation outcomes. Compared to traditional exoskeletons powered by electric motors or hydraulic cylinders, pneumatic muscle actuators (PMAs) offer advantages such as lightweight design, low power consumption, flexibility, and adaptability. This study aims to develop a pneumatic forearm exoskeleton system (PFES) controlled by brainwave signals to provide active training rehabilitation for paralyzed patients. However, the nonlinear hysteresis in PMAs, caused by friction, material elasticity, and internal air pressure dynamics, complicates the design of a unified optimal controller for both inflation and deflation phases. To address this, we employed dual-path fuzzy neural networks (DFNN) to optimize the forearm lifting and lowering functions of the PFES. The DFNN controller combines the strengths of fuzzy logic controllers (FLCs) in handling imprecision and ambiguity with the learning capability and flexibility of neural networks. Compared to a single FNN controller (SFNN), the DFNN exhibited superior performance, achieving significantly lower errors during cycling tests under different loads (1.23 ± 0.09 vs. 3.26 ± 0.16 for a 1&#xa0;kg load and 1.78 ± 0.13 vs. 3.39 ± 0.17 for a 2&#xa0;kg load). The DFNN-controlled brainwave-driven PFES empowers paralyzed patients to actively operate the system through conscious intention, paving the way for more effective active rehabilitation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Designing an EEG Signal-Driven Dual-Path Fuzzy Neural Network-Controlled Pneumatic Exoskeleton for Upper Limb Rehabilitation

  • 期刊论文

  • 发表日期：

    2025年4月9日

查看全文
  • Hao-Hsuan Hsu,
  • Wei-Keung Lee,
  • Po-Lei Lee

摘要

Improving upper limb motor deficits in severely disabled patients is a key therapeutic goal in rehabilitation. Recent clinical studies have shown that upper limb exoskeleton robots can facilitate repetitive training for stroke patients, aiding in regaining movement control and improving rehabilitation outcomes. Compared to traditional exoskeletons powered by electric motors or hydraulic cylinders, pneumatic muscle actuators (PMAs) offer advantages such as lightweight design, low power consumption, flexibility, and adaptability. This study aims to develop a pneumatic forearm exoskeleton system (PFES) controlled by brainwave signals to provide active training rehabilitation for paralyzed patients. However, the nonlinear hysteresis in PMAs, caused by friction, material elasticity, and internal air pressure dynamics, complicates the design of a unified optimal controller for both inflation and deflation phases. To address this, we employed dual-path fuzzy neural networks (DFNN) to optimize the forearm lifting and lowering functions of the PFES. The DFNN controller combines the strengths of fuzzy logic controllers (FLCs) in handling imprecision and ambiguity with the learning capability and flexibility of neural networks. Compared to a single FNN controller (SFNN), the DFNN exhibited superior performance, achieving significantly lower errors during cycling tests under different loads (1.23 ± 0.09 vs. 3.26 ± 0.16 for a 1 kg load and 1.78 ± 0.13 vs. 3.39 ± 0.17 for a 2 kg load). The DFNN-controlled brainwave-driven PFES empowers paralyzed patients to actively operate the system through conscious intention, paving the way for more effective active rehabilitation.