Research On Emotion Recognition Of EMG Signal Method For Exercise Rehabilitation

In robot assisted rehabilitation training process,with active participation has important influence on the effect of rehabilitation training of patients with existing active participation in the rehabilitation process mainly in biomechanical and bioelectrical movement level,does not involve the active participation of the psychological level.In order to solve this problem,the facial emotion recognition method for facial rehabilitation is studied in this paper:First,the design of virtual training scenes.The object of virtual rehabilitation training is the patients in the middle stage of rehabilitation.The proportion of the elderly is large,so the design of virtual scene should be simple and clear,easy to operate,and bring patients a sense of immersion.The training scene was designed with three different levels of virtual rehabilitation training with different levels of emotion,such as "excitement","frustration" and "boredom".Secondly,the acquisition of facial EMG signal.The EMG signal acquisition equipment with high frequency was used to collect EMG signals of the two muscles of zygomatic muscle and frown muscle in the corresponding emotion,and the experimental program stored the performance data under different emotions.At the end of the experiment,the subjects were given a questionnaire about their emotional status.Thirdly,feature extraction of facial EMG signals.The EMG characteristic values were extracted from the two dimensions of time domain and frequency domain analysis of the questionnaire,significant differences in facial EMG feature different training task value,questionnaire survey and different training tasks in investigation and facial EMG feature value,questionnaire and data table,data correlation and facial EMG eigenvalue analysis and that can reflect the physiological characteristics and performance data of the target’s emotional state.Finally,the application of BP network in experiments is presented.The experiment uses gradient descent method,BP network and BP network based on LM algorithm,and chooses BP network as emotion classifier to classify the three kinds of emotion effectively.The classification results show that the BP network based on LM algorithm has better classification recognition effect.