| Due to the current situation of my country's aging population,the incidence of stroke is increasing year by year,so the rehabilitation treatment of stroke population is particularly important.Clinical medical research has proved that positive emotional states can help rehabilitation treatment of stroke patients.At present,the existing rehabilitation robot mainly realizes interactive control through force feedback,and its monotonous repetitive training actions fail to achieve the expected clinical rehabilitation effect.In the subsequent development of rehabilitation robots,how to arouse the enthusiasm of patients,from the psychological level of patients,to ensure the positive emotions of patients during the rehabilitation training process is a shortcut for rehabilitation robots to enter clinical applications faster.For the above reasons,this paper carried out a research on the robot-assisted rehabilitation human-computer interaction control method based on emotion recognition.The specific work content is as follows:First,the background knowledge related to stroke disease is briefly introduced.In view of the shortcomings of traditional rehabilitation training methods,a robot-assisted rehabilitation training system based on emotion recognition is proposed.At the same time,the research status and development of robot-assisted rehabilitation training at home and abroad are discussed.trend.Secondly,a virtual rehabilitation training experiment for stroke patients was designed in combination with games.The virtual rehabilitation training experiment scene was designed using Unity3 D technology,a virtual rehabilitation training simulation platform was constructed,and physiological signals(respiratory,pulse,muscle)of different subjects and normal subjects were taken.Electricity,electrocardiogram,skin electricity),record the performance data during the experiment,and conduct a questionnaire survey on the subjects after the experiment.Perform feature extraction on the collected physiological signals.Thirdly,perform feature analysis on the extracted features,use statistical analysis to evaluate the effectiveness of emotional induction,and verify the differences between physiological signals and performance data of different training tasks and the relationship between physiological signals and performance data and the self-report questionnaire.Correlation.Finally,design an intelligent rehabilitation training system based on emotion recognition,introduce the emotion adjustment mechanism,train the emotion classifier through the four common machine learning algorithms of multi-layer perceptron,support vector machine,K-nearest neighbor,and linear discriminant,support vector machine algorithm The highest recognition accuracy rate was 86.5%.The support vector machine algorithm was used to cross-validate the data obtained from healthy people,patients' healthy side and patients' affected side.Comparing the effect of human-computer interactive rehabilitation based on emotion recognition with the effect of human-computer interactive rehabilitation based on performance data,it is found that the task selection of the robot-assisted rehabilitation training system based on emotion recognition is more challenging,and the task difficulty of the system is different from that of patients.The degree of subjective willingness is up to 85%,which can ensure that patients can complete the rehabilitation training tasks suitable for their conditions with positive emotions during rehabilitation training,which helps to improve the efficiency of rehabilitation. |
