Preliminary Study On Cognitive Load Of Flight Simulation Training Based On EEG

With the rapid development of modern industrial technology,the human-machine system is becoming more and more complex.Due to the limited ability of human brain to receive and process information,cognitive activities in human-machine system will bring cognitive load to the brain.Appropriate cognitive load is helpful to improve learning efficiency,but too high cognitive load will backfire.The study of cognitive load in flight simulation training can guide the training of pilot cadets effectively.EEG is widely used in cognitive load assessment because of its advantages such as low cost and portable device.Approximate entropy(ApEn),wavelet packet transform and Hilbert-Huang transform algorithms are used to analyze the EEG from different angles in the simulated flight in this paper,and the analysis results could reflect the cognitive load of cadets.Firstly,basic knowledge and experiment involved were introduced in this paper.The basic knowledge about EEG is explained from the microcosmic formation mechanism,basic rhythm and physiological characteristics of EEG,and the functional zoning of the brain was described briefly.Meanwhile,the mechanism of cognitive activity,the origin and development of cognitive load theory,and the main viewpoint were introduced.The experiment involved is a conventional airfield traffic pattern simulation flight.The subjects were all pilot cadets from different learning stages,and the flight scenes include clear weather and rain patterns.The EEG data collected covered the whole flight simulation.Secondly,the ApEn algorithm which as the core and the statistical methods which as the auxiliary is used to analysis the brain electrical signals.The variation characteristics of ApEn in the flight are summarized.The difference of ApEn was pointed out from the weather,genders and training situations.Further,it was observed that the ApEn increases gradually with the training time,and the approximate upper bound threshold of ApEn is obtained.It is found that ApEn is an effective algorithm for EEG data analysis because of its advantage in processing random signals.The results show that ApEn can reflect the cognitive load to some extent.Finally,the wavelet packet decomposition and Hilbert-Huang transform were used to analyze in the time-frequency aspect.The variation characteristics of energy of each flight phase were summarized,and the regularities of energy distribution of each phase based on large samples were obtained with the statistical charts,which validation by support vector machines.In addition,a basic method of quantitative assessment of cognitive load was proposed,whose change process was shown based on advantage of empirical mode decomposition.