Fatigue State Recognition Of Miners Based On EEG And ECG Fusion

With the rapid development of the economy and the mechanization of mining,there has been an increase in work-related accidents,posing a threat to the safety and lives of miners.The main types of accidents in mining production are object impact and mechanical injuries,with human error caused by fatigue being the main reason for accidents.Therefore,the identification of a miner’s fatigue state is of great theoretical and practical significance to ensure their safety and prevent accidents.However,research on this topic is relatively scarce,and it is necessary to explore a fatigue state identification and monitoring method that is suitable for the mining industry in China,based on the fusion of multiple physiological characteristics.This paper proposes a fatigue state identification method based on PCA-SSA-BP,which integrates the fatigue characteristics of both brain waves and electrocardiograms of miners,in order to improve the accuracy of fatigue identification.Additionally,the paper conducts a comparative analysis of the model’s application.The main research contents are as follows:(1)Optimization of multiple physiological feature indicators for identifying miner fatigue and construction of a dataset.Through a literature review and analysis,it was found that EEG and ECG signal feature indicators are sensitive to changes in fatigue status.Therefore,EEG and ECG signals were selected as the raw data and after processing,8-dimensional EEG feature indicators and 7-dimensional ECG feature indicators were obtained.To obtain an optimal set of feature indicators,t-tests were performed on the Pearson correlation coefficients between the features.Subsequently,a dataset of miner fatigue samples was constructed.(2)Construction of a miner fatigue recognition model based on PCA-SSA-BP.The feature-level fusion strategy and principal component analysis(PCA)were used to fuse and reduce the dimensionality of the sample dataset,and to determine the fusion indicators of fatigue state features(relative energy ratio of different rhythmic waves F and heart rate variability index SDNN)as inputs for the model.The initial weights and thresholds were optimized using the sparrow search algorithm to construct a miner fatigue recognition model based on the PCA-SSA-BP algorithm.(3)Comparison and analysis of the application of the miner fatigue state recognition models.Empirical evidence demonstrates the effectiveness of the PCA-SSA-BP model for fatigue recognition.In this study,we established the BP neural network model and GA-BP neural network model for comparison.The models were compared based on four metrics: precision,recall,F1 score,and accuracy.The results show that the neural network model based on PCA-SSA-BP is more ideal for miner fatigue state recognition,with a recognition accuracy of 93.75%.Finally,suggestions and countermeasures are proposed for the safety management measures of mining machinery equipment operation,monitoring and early warning of fatigue status.Through the analysis of physiological characteristics of miners during fatigue work,this paper proposes a fatigue state recognition method based on PCA-SSA-BP.The research results provide a theoretical basis for the formulation and improvement of rest regulations for enterprise employees,and have reference value for the development of related portable fatigue state monitoring hardware.