Research On Fatigue Recognition Of Deck Officers Based On EEG Signals

The study indicates that approximately 80% of marine accidents involve human factors,with fatigue being a significant contributing factor to human errors.The exhaustion of deck officers in the shipping industry is widespread,and accidents caused by fatigued driving among deck officers can cause significant casualties and property damage.Monitoring the fatigue levels of deck officers and taking corresponding preventive measures can significantly reduce the incidence of maritime accidents,thus having significant economic and social value.In recent years,the development of fatigue detectors for drivers and pilots based on electroencephalogram(EEG)technology has made it possible to quickly and accurately monitor the fatigue level of deck officers.This article analyzes the EEG monitoring methods for driver fatigue in various industries,and organizes a simulated ship driving experiment with 21 subjects to establish a three-level fatigue recognition method for deck officers based on EEG,including "alertness","moderate fatigue",and "fatigue".The experiment shows that the detection accuracy of EEG signals is reduced due to individual differences and reduced sensitivity,so in order to improve the performance of fatigue classification methods based on EEG signals,this article adopts methods such as EEG fusion features,Bi-directional Gated Recurrent Unit(BiGRU)deep neural network classifiers,and the ADTIDO(Automatic Detect the TIred Deck Officers)EEG channel integration algorithm to improve classification efficiency.The results show that,based on the experimental design and data processing in this study,the classification accuracy of the Bi-GRU classifier reached 90.19%.The proposed fusion feature improves the classification accuracy by 1.89% compared to using a single EEG feature as input,and also has certain advantages compared to the fatigue features and classification algorithms used by other scholars.Based on this,the ADTIDO channel integration algorithm was proposed,and the classification accuracy of deck officers’ fatigue level reached 95.74%.Current research indicates that this method can achieve significant effects in the classification of three levels of fatigue across individual deck officers,making it an effective method for objectively monitoring their fatigue.Based on this,operational recommendations or requirements can be proposed for deck officers of different fatigue levels.