Research On Fatigue State Detection Method Based On Electroencephalography Signal

With the advancement of modern technology and the tremendous changes in lifestyles,people enjoy the overall improvement of work and life efficiency,and at the same time become more and more work pressure,it is easy to break through the limits of one’s endurance,and then lead to physical and mental fatigue;People in the state of work and study efficiency decline,a long-term fatigue state may even cause greater danger.Real-time and accurate detection of human fatigue,reminding people to stop work appropriately and rest,not only to better complete the next work,but also to be responsible for the society and other people.This is an important research topic in the current scientific field.EEG signals contain a large amount of human-related information,which is an important way to detect human fatigue.In this paper,a single-channel EEG signal acquisition device based on experimental design is used to collect related EEG signals.After filtering processing,feature values are extracted and graded and evaluated to achieve the purpose of facilitating and quickly determining the fatigue state of the human body.The specific work is as follows:First of all,the upper computer of the acquisition experiment is written based on the C++ language,and the relevant EEG signals are collected with special equipment.In signal processing,this article focuses on the current situation of single-channel EEG signals with a lot of noise and difficulty in removing artifacts.A comprehensive algorithm combining independent component analysis and variational modal decomposition is used to focus on EEG signals.Removal algorithm of electrooculogram artifacts.The comprehensive algorithm has been actually tested and compared with the results,which can avoid the aliasing phenomenon of the frequency band that may occur in the empirical mode decomposition of the signal,and the artifact removal effect is excellent.Then,according to the characteristics of high complexity of EEG signal and fluctuation of complexity in the two states of fatigue and awake,we chose fuzzy entropy as a means to extract EEG signal features.The entropy value is usually proportional to the amount of information contained in the signal or system.According to data,normal humans have more abundant physiological signals in the waking state,and the entropy value of the EEG signal may be larger,while in the fatigue state,it is relatively simple,and the inferred entropy value may be smaller.Therefore,this paper uses the entropy value to measure the brain The complexity of electrical signals and the judgment of fatigue.Finally,in order to obtain an objective and quantified classification and evaluation of fatigue status,this paper decides to use support vector machine as the classifier used in this paper to classify and recognize the extracted EEG signal feature indicators.After the feature extraction of the EEG signal,the extracted fuzzy entropy value of the EEG signal of the test volunteer is used as the feature vector,and the four levels of the fatigue state of the test volunteer are classified and identified through the support vector machine,and the experimental data The samples are divided into two groups: test set and training set.The subjective evaluation and statistical analysis of the fatigue state of the volunteers are comprehensively compared,and satisfactory test results are obtained,verify the feasibility of the EEG fatigue classification experiment based on support vector machine.