以直接/間接腦波人機介面為基礎之中風病人復健輔具系統-總計畫暨子計畫一：生物電磁刺激裝置開發與肌電輔具系統(3/3)

Improving motor deficits in the upper limbs post-stroke is a primary goal of rehabilitation, as approximately 85% of patients cannot spontaneously recover upper limb function, necessitating effective rehabilitation methods. Our research developed an EEG-driven exoskeleton rehabilitation system, utilizing Pneumatic Muscle Actuators (PMA) to assist in upper limb rehabilitation. Compared to traditional motors, its high power output, low cost, and lightweight nature make the entire rehabilitation system more efficient and economical. We integrated a Fuzzy Neural Network (FNN) for control within the system, eliminating the need for complex modeling and achieving precise control performance. By combining Motor Imagery (MI) with the rehabilitation system, we offer a more comprehensive rehabilitation plan, enhancing patient engagement and recovery potential. Experimental results demonstrate the system's effectiveness in achieving accurate trajectory tracking and control, even under load conditions. Our research indicates that this system is very effective in assisting patients in regaining motor abilities through imagination-based rehabilitation exercises. This project developed a brainwave recognition neural network, which interprets EEG signals to identify brainwave intentions, and designed transcranial and muscular electrical stimulation for rehabilitation.