Research On The Evaluation Of Virtual Force Feedback In Cross Subjects Based On Multi-Mode Micro Current Stimulation

Virtual reality technology can highly restore human visual and auditory perception behavior in the real world,but the virtual reality scene which only contains audio-visual interaction can not meet the growing needs of consumers,such as the sense of reality and immersion,and virtual reality multi-mode interaction is more and more valued.The research of force feedback interaction mode and functional micro current stimulation in virtual reality interaction provides theoretical basis and practical feasibility for virtual reality force feedback interaction.At present,in the field of virtual reality power feedback based on micro current stimulation,the following problems need to be studied.Firstly,there are few researches on multi-mode research and experimental verification of micro current for virtual force feedback,and the cross subject multi-mode micro current stimulation force feedback correlation data set also needs to be constructed;secondly,there are few researches on the removal of personalized differences among different subjects in the field of virtual force feedback,and there is insufficient research on the influence of personalized characteristic parameters on force feedback;finally,there is a lack of research on the effect of personalized characteristic parameters on force feedback cross subject force feedback evaluation model for multi-mode micro current stimulation.In view of the above problems,the main work of this paper is as follows:(1)Research on micro current stimulation mode: modulation of triangular wave,selection of electrical stimulation mode and pulse width parameters.Through the investigation of human epidermis reality perception and the threshold and peak value experiment of electrical stimulation current,the triangular wave micro current with amplitude of 10 m A,15 m A,20 m A,25 m A,30 m A and 35 m A was determined,and the duration of each amplitude current was 30 s.(2)Research on experimental paradigm and data set construction of cross subject virtual force feedback data acquisition under multi-mode micro current stimulation: including the formulation of experimental paradigm,cross subject physiological parameters acquisition and force feedback data segmentation and processing.Based on the experiment of micro current stimulation,according to the degree of muscle fatigue of the subjects,the experimental paradigm of one minute rest between each amplitude of current stimulation and the optimal electrode position for biceps brachii were determined.The force feedback data of each subject was recorded in real time by dynamometer recording software,and the continuous current value and force feedback value were segmented synchronously.In this experiment,a total of20 subjects' data and 8 floating-point physiological parameters were collected,and 2207 cross subject force feedback data sets were obtained.(3)The multi-mode micro current stimulation virtual force feedback Association Evaluation Model Based on individual global effectiveness features is studied: cross subject physiological feature extraction and force feedback evaluation model based on global features.Feature engineering is to extract AC*height,AC*BFP*height,(AC + BFP + height)/3 and AC*BFP*height based on cross subject physiological parameters through feature importance,correlation analysis and feature combination methods.And through the XGBoost,LGBoost,GBDT,random forest,decision tree model and model fusion to fit the force feedback data,it is verified that the fitting effect of XGBoost model and GBDT single model on the force feedback data is better,the average absolute error of XGBoost is 2.21,and the correlation coefficient between the force feedback prediction value and the test value of gbdt is 0.9;while in the aspect of model fusion,XGBoost and GBDT single model have better fitting effect on the force feedback data The fitting correlation coefficient of random forest fusion model to force feedback value is 0.9.Based on micro current stimulation and machine learning technology,this paper studies cross subject virtual force feedback from a new perspective,and provides a force feedback evaluation model based on the effective global characteristics of cross subjects,which enhances the immersion and realism of virtual reality feedback interaction,and provides a new idea for cross subject virtual reality feedback research.