一种基于卷积神经网络的下肢动作模式识别方法

针对目前下肢动作模式识别技术存在的数据量少、识别率低的问题，提出了一种新的基于卷积神经网络的下肢动作模式识别方法。以下肢步态动作识别为对象，采集无负重平地行走，无负重上/下楼及负重上/下楼5种步态的表面肌电信号（surface electromyography，sEMG），对sEMG进行特征提取，构建了一种以特征集作为输入的卷积神经网络，并比较了其与另外几种传统分类识别方法的识别准确率和工作特征。实验结果表明，新方法对于5种步态的平均识别准确率大于95%，错误率都低于8%，具有较高的准确性。因此所提方法的输入特征集更能代表预测模型特征，模式识别率更高，可为康复医疗机器人、助力机器人等设备改善下肢运动功能提供参考。

A pattern recognition method of lower limb movements based on convolutional neural network

In order to solve the problems of low data and low recognition rate in the current pattern recognition technology of lower limb movements,a new lower limb movement pattern recognition method based on convolutional neural network was proposed.The lower limb gait movements recognition was taken as the object,and the surface electromyography (sEMG) signals of five gaits of walking on flat ground without weight,going up/down stairs without weight,and going up/down stairs with weight were collected.Based on the feature extraction of sEMG,a convolutional neural network with feature set as input was constructed,and the recognition accuracy and working characteristics of several other classification and recognition methods were compared.The experimental results show that the average recognition accuracy of this method for five gaits is greater than 95%,and the error rate is less than 8%,which has high accuracy.The input feature set of the method can better represent the characteristics of the prediction model,and the pattern recognition rate is higher,which provides some reference for the improvement of lower limb motor function of rehabilitation medical robots,power-assisted robots and other equipment.