Feature Extraction And Classification Recognition Of Operator’s Mental Workload In The Teleoperation System

Early research on the teleoperation system mostly focused on the system ontology,with the synchronization between master control commands and remote robot status as the main research content.However,with the continuous deepening of human-machine interaction research,operators’ characteristics have become the decisive factor in the task completion of the teleoperation system.As the typical human-machine interaction system,there are many factors in the teleoperation system that will affect operators’mental workload.Among these influencing factors,communication time-delay as the decisive factor for the stability and transparency of the teleoperation system is more worthy of attention.As a consequence,this thesis explores the influence of the change of communication time-delay in the teleoperation system on operators’ mental workload state from the perspective of data mining and classifies and evaluates operators’ mental workload state.This thesis extracts the characteristics of the mental workload from subjective measurement,work measurement,and physiological measurement.In subjective measurement,the NASA-TLX scale is used to analyze operators’ feelings about their functional status,and it is found that operators need to pay more energy in an operating environment with larger communication time-delay to achieve the same operational performance as a lower time-delay environment.In work measurement,the effective operation interval ratio is proposed as a cognitive task performance indicator.It is found that as the communication time-delay continues to increase,operators’ logical consciousness during the execution of the task gradually becomes blurred,and the degree of their panic has increased.In physiological measurement,the EEG signal with higher time resolution is selected for analysis.By calculating the brain electrical energy of the operator to draw the brain-associated topographic map,it is verified that the logical reasoning ability of the frontal lobe brain area was more needed by operators in the delayed environment.At the same time,the ratio of high frequency to low frequency in the EEG power spectrum is gradually increasing,which means that operators’ working ability has declined.In addition to the power spectrum,sample entropy is also a characteristic indicator commonly used in EEG signal analysis.It turns out that as the communication time-delay increases,the complexity of the EEG signal is increasing,while the stability is gradually decreasing.In summary,according to the changes of mental workload characteristic indicators,it is confirmed that the increase of the communication time-delay in the teleoperation system will make operators’ mental workload state worse.Besides,based on the problem that the single feature indicator may not be able to accurately evaluate the state of mental workload,this thesis uses principal component analysis to select various indicators to construct a new comprehensive score index.After the training and classification of the support vector machine,the correct rate of the score indicator is as high as 92%,which can be used to evaluate operators’ mental workload.From the perspective of ergonomics,this thesis reveals the internal connection between the communication time-delay in the teleoperation system and the mental workload of the operator.Moreover,the research results are also applicable to selecting excellent teleoperators,explaining the mechanism of human brain processing time-delay,and evaluating the effectiveness of system interface design.