Non-contact Objective Detection Of Brain Cognitive Load Based On Distributed Radar Breathing Signals

With the rapid development of informatization of the society,humans face increasing mental workload,which brings more cognitive load.Excessive cognitive load will not only reduce productivity,but may also lead to mistakes in decision-making.Therefore,detection of cognitive load is of great significance.Detection of cognitive load based on physiological signals is objective and accurate.Currently physiological detection methods of cognitive load obtain physiological signals through contact devices,which may cause discomfort to the subjects.Bio-radar can obtain the physiological signal of respiration in a non-contact manner,but existing researches are conducted on the premise that the human body faces the radar.In actual scenarios,the orientation of the human body may change at any time,which degrades the waveform quality of the respiratory signal and ultimately affects the measurement accuracy.Aiming at the above problems,this paper proposes a non-contact objective detection method of cognitive load based on respiratory signals obtained by distributed radar.The main research work and contributions are as follows:(1)In terms of data acquisition: Targeting at the problem of degraded waveform quality of the respiratory signal due to direction changes of the human body,this paper builds a distributed radar data acquisition platform to acquire radar echo signals in multiple directions and multiple channels.The concrete work includes hardware implementation,host computer design,and arrangement design of distributed radar.In addition,this paper uses the experimental paradigm of mental arithmetic task to induce high and low cognitive load levels.To be close to the actual scene,the subjects completed each stage of the experiment with their body facing different directions.This paper verifies the validity of the experiment by analyzing task performance and the subjective scale.(2)In terms of data processing: This paper uses the demodulation algorithm to demodulate the radar echo signal,including I/Q channel calibration and arctangent demodulation,and verifies the reliability of the algorithm by simulation.To solve the problem of how to choose the optimal signal from multiple demodulation signals,this paper proposes an adaptive channel screening method based on waveform quality.The method first eliminates channels with abnormal waveform through body motion interference detection and periodic detection,then selects the channel with the highest signal-to-noise ratio as the final output channel based on the principle of optimal signal-to-noise ratio.(3)In terms of detection of cognitive load: This paper extracts the time domain and frequency domain features of the respiratory signals,then eliminates individual differences in extracted features,and finally uses sequential backward selection algorithm to screen features.This paper builds feature data set,and then detect high cognitive load and low cognitive load using SVM and KNN models.The experimental results show that SVM and KNN achieved the highest classification accuracy of 87.03% and 81.1%,respectively.SVM achieved better performance than KNN,and under the highest classification accuracy its recognition rate for high cognitive load and low cognitive load is 90.23% and 84.17%,respectively.