| Psychological stress is a serious social public health problem faced by all countries in the world,which is directly related to people’s health and social happiness.The pulse signal has the advantages of easy collection,low cost and easy analysis.However,in the process of using the pulse signal to measure psychological stress,due to noise interference such as motion artifacts,the extraction accuracy of the Heart Rate Variability(HRV)signal is low,and the extraction time is also long.In addition,due to the lack of effective analysis methods,it is difficult to count HRV characteristics in a short period of time,which will lead to the loss of some important information representing psychological stress.This has led to difficulties in detecting individual differences in psychological stress and making it difficult to quantify accurately.It has caused difficulties such as the difficulty in detecting individual differences in psychological stress and the difficulty in accurately quantifying it.In order to solve the above problems,the thesis conducts research on pulse signal acquisition,accurate extraction of HRV signals,psychological stress assessment and other issues.The research results not only provide a new method for psychological stress detection,but also can be widely used in the diagnosis of mental health diseases,evaluation of rehabilitation effects,research on the heart and brain system,etc.It has important research significance and value.The main work and innovations of the thesis are as follows:(1)A method for acquiring pulse HRV signals based on the Variational Mode Extraction(VME)algorithm is proposed.First,based on the dynamic model of two Gaussian functions,the pulse signal is simulated,and then the relationship between different types of noise in the pulse and the penalty coefficient value(alpha)of the VME algorithm is studied.The improved VME algorithm is also proposed to address the problem that baseline drift noise in the VME algorithm can lead to reduced HRV extraction accuracy.The research results show that compared with the existing modal decomposition method,this method requires fewer parameters to be configured,and the HRV signal acquisition speed and accuracy have been significantly improved.Compared with the variational modal decomposition method,the accuracy is increased by 40%,and the time is saved by 40 times.It provides a basis for extracting psychological stress features from HRV signals.(2)A heartbeat mode decomposition method that can be used for HRV analysis is proposed,and a new feature bmNN(beat mode of Normal-to-Normal intervals)is generated for quantitative detection of psychological stress levels.The method is based on the heart rhythm consistency hypothesis theory.First,quantify the time structure characteristics of the HRV signal sequence,count the occurrence frequency of adjacent heartbeat interval changes under the modal scale m,and calculate the eigenvalues(bmNN)corresponding to different heartbeat modes bm(beat mode,2m in total)in the HRV sequence.Then,the nearest neighbor classification algorithm is used to classify the target stress data,and the classification accuracy rate under different bmNN values is analyzed,and then the relationship model between bmNN features and psychological stress is constructed.Finally,the optimal heartbeat mode bm,which characterises the change in stress,is selected to achieve differential detection of psychological stress.And it was also tested on two psychological stress datasets,car driving and Montreal,and the results showed that the identification accuracy of the method in psychological stress dichotomization was 93.7%and 96.3%,respectively.Compared with the traditional HRV time-domain analysis method,the heartbeat mode decomposition method utilizes the time structure characteristics,and by adapting the optimal m and bm values,it can realize the quantitative evaluation of different psychological states and the detection of individual differences.It is an important supplement to the traditional method.It also suggests that the bmNN trait can be a valid measure of psychological stress levels and has a wide range of applications in other quantitative assessments of health.(3)An experimental study of a method for detecting psychological stress based on pulse signals was carried out.First developed a polyvinylidene fluoride(PVDF)piezoelectric pulse sensor.Then the experimental environment of the Montreal psychological stress task was built.In this environment,the pulse signal stress data acquisition,HRV signal extraction,and HRV data analysis based on the heartbeat mode decomposition method have been completed.Finally,the effects of individual differences and spatiotemporal variations on the detection of psychological stress are investigated.The results of the study show that:Adapting different optimal bm for each experimental data of different individuals,it can detect the psychological stress of different individuals in the current experimental time and space,and the average accuracy rate of recognition in stress dichotomy is 96.3%;adapting the same optimal bm for all experimental data of the same individual,it can detect the psychological stress of the individual in different experimental time and space,and the average accuracy rate of recognition in stress dichotomy is 93.6%.It proves that the heartbeat mode decomposition method proposed in this thesis can not only detect the psychological stress of different individuals in the current time and space,but also detect the psychological stress of the same individual in different time and space. |
