Research On Operational Efficiency Decline Method Based On Multi-dimensional Information Classification And Fusion

Nowadays there are many workplaces such as aerospace and deep sea voyages that require a full mental state of work.If the operator's attention is distracted or his reaction is slow,serious accidents may occur and cause very adverse social effects.The focus of current safety work in our country are enhancing safety technology research and development,establishing a safety technology research and development system,and promoting the research and development of major general key technologies and equipment.Therefore,based on the above,this paper aims to investigate the operational efficiency decline detection technology for operators in high-risk industries.The paper is based on the experiments carried out in a submarine cabin.Firstly,it introduces the background and significance of the research.A review and analysis of the current research status of physiological characteristics such as Electroencephalogram(EEG)and Electrocardiogram(ECG)data,multi-dimensional information fusion and the operational efficiency decline,and an overall research framework is formed accordingly.Secondly,an analytical model of the a priori probability of operational efficiency decline is developed.On the basis of quantitative analysis performed on the factors influencing the decline in operational efficiency and a weighted evaluation carried out on these factors,an a priori probability analysis of the decline in operational effectiveness was established on the above quantified impact indicators combined with the weighted values using ordinal relationship analysis.Next,the multidimensional information collected during the experiment were pre-processed and analysed by feature extraction.In the process of pre-processing and cleaning the EEG data,the proposed independent component analysis and automatic discrimination method were used to remove artefacts from the EEG,which not only effectively removes artefacts but also saves a lot of time and labour costs.The power spectral density and energy characteristics are extracted from the EEG data for different bands.Both R-R interval features and the energy features on the ECG waveform were extracted to fully reflect the information contained in the ECG.Afterwards,a multidimensional information classification model and a multidimensional information fusion model were constructed for operational efficiency decline.In the classification of each physiological data,support vector machine optimized by genetic algorithm was to improve the related parameters selection and further improve the model performance and classification accuracy.The performance improvement of the classifier is verified by comparing the classification accuracy of EEG data and ECG data with traditional support vector machines.A fuzzy logic inference system was used to fuse the physiological data collected in the experiments with the task data characteristics and the subjective scale with the work time in the experimental arrangement to derive the final operational efficiency decline values.The feasibility and accuracy of the multidimensional information classification and fusion model of operational effectiveness decline were verified by calculating the error indicators comparing the results of the single dimension before fusion and the results after fusion.Finally,the work done in the study of operational efficiency decline is summarised and analysed,and future directions for the research of operational efficiency decline technology are proposed.