基于多传感器信息的新型穿戴式上肢外骨骼康复机器人

A new wearable upper limb exoskeleton rehabilitation robot based on multi-sensor information

目的设计基于多传感器信息的新型穿戴式上肢外骨骼康复机器人,以解决上肢外骨骼康复机器人便携性不佳、患者参与度较低、训练模式自适应不足等问题,并探究受试者穿戴外骨骼时肌肉激活程度、肌电信号预测关节角度的准确性以及实现上肢康复训练的可行性。方法该设备机械结构包括肘关节和腕关节,采用模块化设计并结合3D打印技术;控制系统包括肌电采集、应力采集、姿态采集等单元,并设计主动、被动和助动三种训练模式。受试者穿戴外骨骼机器人后进行屈-伸肘实验,对比有、无辅助力时手臂肌肉激活程度;分析肘关节角度,并对比肌电信号预测的关节运动角度;验证机器人运行性能与应力检测效果。结果受试者穿戴外骨骼康复机器人安全可靠地完成了屈-伸肘动作,受试者肱二头肌、肱三头肌肌肉激活程度在有、无辅助力时分别减弱约32%、11%,肌电信号预测关节角度准确度约95%,应力测量值误差均低于5%。结论上肢外骨骼机器人可以给人体提供辅助力、预测关节角度,机器人通过肌电、应力以及位置信息辅助患者实现上肢康复训练具有可行性。

Objective To design new wearable upper limb exoskeleton rehabilitation robots based on multi-sensor information,in order to solve problems of poor portability of upper limb exoskeleton rehabilitation robots,the low degree of participation for patients and insufficient training mode adaptability,and to explore the muscle activation degree of participants wearing exoskeletons,accuracy of electromyographic signal predicting the joint angle and the feasibility of the upper limb rehabilitation training.Methods The mechanical structure of the device,which included elbow and wrist joints,was modular and combines 3D printing technology;the control system included electromyography(EMG)acquisition,stress acquisition,attitude acquisition and other units,and three training modes of active,passive and assisted were designed.After the subjects wore the exoskeleton robot,the elbow flexion and extension experiment was conducted to compare the activation degree of arm muscles with and without auxiliary force;the elbow angle was analyzed and the joint motion angle predicted by EMG signal was compared.The running performance and stress detection effect of robots were verified.Results The subjects with the exoskeleton rehabilitation robot completed the flexion and elbow extension movement safely and reliably.The muscle activation degrees of the biceps and triceps of the subjects were reduced by 32%and 11%with and without assistance,respectively.The accuracy of EMG signal in predicting the joint angle was about 95%,and the error of stress measurement was less than 5%.Conclusions The upper extremity exoskeleton robot can provide human bodies with auxiliary force and predict the joint angle.It is feasible for the robot to assist patients to achieve upper extremity rehabilitation training with EMG,stress and position information.