利用运动准备电位的人体下肢自主运动意图预先感知方法

A Preperception Method for Voluntary Movement Intention of Lower Limb Using Readiness Potential

针对下肢外骨骼机器人穿戴者的运动与机器人本身响应之间存在时间迟滞而导致控制实时性差的问题,提出了一种利用运动准备电位(RP)的人体下肢自主运动意图预先感知方法。首先,根据人体下肢运动的脑电(EEG)产生机理和耦合神经元集群模型仿真了下肢产生运动意图时大脑运动皮质区的RP响应,验证了下肢产生运动意图时其EEG中存在RP;其次,利用下肢肌电(EMG)确定了下肢自主运动的起始时刻并完成EEG中自主下肢运动事件的触发标记,同时利用多元经验模式分解(MEMD)方法去除了EEG中的运动伪迹成分,进而利用叠加平均法和模板匹配法实现了下肢自主运动的RP提取及其运动意图的预先感知;最后,搭建检测系统进行实验验证。实验结果表明:该方法可在离线情况下于人体下肢自主运动开始前约120 ms实现下肢运动意图的预先感知,其平均识别正确率为74.4%。该方法实现了人体下肢自主运动意图的预先感知,可为外骨骼机器人的柔顺控制提供信息提前量。

A preperception method for voluntary movement intention of lower limb using readiness potential(RP)is proposed to solve the problem of poor real-time control caused by the time lag between human movement and robot’s response in lower limb exoskeleton robot system.Firstly,the electroencephalogram(EEG)signals of motor cortex are simulated based on EEG generation mechanism of human lower limb movement and a coupled neural mass model.It is verified that the RP does exist in EEG of motor cortex during lower limb voluntary movement.Secondly,the electromyography(EMG)of lower limb is used to detect the onset time of lower limb voluntary movement,and to complete the trigger marker for voluntary movement of lower limb in EEG signals.Meanwhile,motion artifacts in EEG signals are removed by using the multivariate empirical mode decomposition(MEMD).Furthermore,RP extraction and preperception of lower limb voluntary movement intention are realized by a superposed average method and a template matching method,respectively.A detection system is finally built to perform experimental verification.Experimental results show that the proposed method achieves the lower limb voluntary movement intention 120 ms in advance,and the average accuracy is 74.4%.It is concluded that the proposed method realizes the precognition of lower limb voluntary movement intention,and provides advance information for compliant control of lower limb exoskeleton robot.