基于脑肌电反馈的虚拟康复训练系统设计

Design of virtual rehabilitation training system based on EEG and sEMG feedback

针对虚拟康复系统在个体适应性、安全可行性和主动参与性等方面的研究需求,提出一种基于脑肌电反馈的虚拟康复系统。采集执行不同手势动作对应的脑电信号及表面肌电信号,提取不同肌肉模块的肌电特征送入支持向量机模型进行运动意图识别;提取脑电和肌电疲劳特征,并提出一种萤火虫-模糊神经网络算法,通过脑肌电疲劳特征实时优化调节虚拟场景的控制参数。最后,搭建包含虚拟场景及反馈控制策略的虚拟康复系统,并针对上肢肘关节屈伸、肩关节前屈后伸动作进行康复训练实验,基于肌电特征模式识别结果实现对虚拟场景及目标的控制,基于脑肌电疲劳特征优化调整场景控制参数,通过虚拟系统康复实验验证系统的有效性。

To satisfy the key requirements for virtual rehabilitation systems,such as limitations in individual adaptability,safety,and active participation,an EEG and sEMG feedback based virtual rehabilitation system is proposed in this study.The system collects the EEG signals and sEMG signals to detect the features of different gestures.Then,the SVM motion intention recognition model is constructed by utilizing sEMG features from different muscle modules as input.An FA-FNN algorithm is proposed to optimize and regulate the scene control parameters in real time based on the detected fatigue feature of EEG and sEMG.Finally,a system with virtual scene and feedback control strategy is built and rehabilitation training experiment with upper limb elbow flexion-extension and shoulder anteflexion-back extension as paradigm is conducted.The virtual scene and object control is realized with sEMG feature pattern recognition results,and the scene control parameters is optimized and regulated with fatigue feature of EEG and sEMG.The effectiveness of system is evaluated by virtual rehabilitation experiment.