基于协同效应超前预测偏瘫患侧肢体运动意图方法的研究

中国脑卒中发病率世界第一，致残率极高，严重影响了患者的运动功能。若能有效康复患者的运动功能，使其回归正常的生活，势必会大大减轻社会和家庭的负担。对于偏瘫患者的康复，目前以被动物理刺激技术为主，缺乏患者主动参与。本项目基于人类肢体运动具有双侧协同的特点：（1）利用健患双侧的脑电（EEG）、表面肌电（sEMG）等运动控制信号，联合肌动图（MMG）、空间轨迹（ST）等运动结果信号作为超前预测系统的输入，通过健侧运动意图信号训练递归神经网络，使用长短期记忆网络超前预测患侧的运动意图；（2）依据实时的健患双侧多模运动信息，推导出双侧主动协同控制算法；结合运动想象和视觉反馈刺激，配合双侧协同康复训练系统开发出人机共融的先进康复方法；（3）采用柔性应力传感器监测评估肌痉挛程度，以实时补偿优化协同超前预测系统。项目完成将为双侧协同的偏瘫康复机器人关键技术以及肢体运动主动康复的个性化策略提供理论基础和依据。

The morbidity of stroke in China is the highest around the world. And the motor function of patients is affected severely because after stroke has a high disability possibility. Thus, the burden of society and family can be decreased significantly by recovering the motor function effectively. Currently, passive physical stimulation is mainstream rehabilitation therapy to recover motor function. However, the patients lack active participation in such therapy. Based on the simultaneous mechanism of the human motor, (1) we plan to utilize the signals from the normal and the impaired side limbs as input of the prediction system, including Electroencephalography (EEG) signal and surface Electromyography (sEMG) signal represented as motor intention, Mechanomyogram (MMG) signal and spatial trajectory (ST) represented as motor result. After the detail segmentation, the high dimensional timing signals will be input into the long short time memory network for predicting the anticipatory motor intention of the impaired limbs. (2) Based on the real-time bilateral multimodal motion signals, the bilateral active simultaneous control algorithm will be deduced; under the stimulation of motor intention and visual feedback, advanced human-computer integration method based on bilateral rehabilitation training could be realized. (3) Meanwhile, a flexible stress sensor is used to monitor and assess the level of muscle spasticity for compensating and optimizing the simultaneous control system. This research could provide the foundation and basis for the key techniques of bilateral simultaneous hemiplegia rehabilitation robot and personalized strategy of active rehabilitation of limb motor.