This article outlines a study on driver mental workload classification based on two categories of information in electroencephalography (EEG) data: event-related potentials (ERPs) and frequency band powers (BPs). For data acquisition, a simulated driving task, the Lane Change Task (LCT), was combined with a secondary auditory task, the Paced Auditory Addition Serial Task (PASAT). Workload levels were manipulated by changing the speed settings and the paces of the PASAT. The ERPs and BPs (including the five frequency bands of delta, theta, alpha, beta, and gamma) were extracted from the recorded EEG data as the input for a powerful classifier, Adaptive Boosting (Adaboost). The ERP based classification showed a mean accuracy of 63 % for the speed induced workload in the single task condition and a mean accuracy of 66 % for the PASAT induced workload in the dual task condition. Conversely, the BP based classification showed a mean accuracy of 85 % for both speed induced workload and PASAT induced workload classification. This indicates that BPs might be a better choice for the evaluation of the driver mental workload. With the new role of supervisory control in advanced humans-machine systems, there is an increasing requirement on the assessment of human mental workload and performance. Introducing electroencephalography (EEG) technology into the man-machine interface provides an effective methodology to evaluate human functional states. In the last decades, characteristic changes in the event-related potentials (ERPs) and the EEG frequency band powers (BPs) that reflect levels of workload have been identified. ERPs are the stereotyped electrophysiological responses to an internal or external stimulus, while BPs are typically extracted by dividing EEG data into frequency bands of delta, theta, alpha, beta and gamma. Other authors identified a change in the amplitude of P300, a positive peak around 300 ms after the stimulus onset, caused by task difficulty. Additionally, several prominent band power features in EEG data have been reported to be sensitive to the variations in mental effort. For instance, a decrease of parietal-occipital alpha rhythm and an increase of frontal theta rhythm have been observed as the mental effort of a task increases. Based on these features, a number of researchers have attempted to classify mental workload using different classifiers. Other papers revealed an accuracy of 80 % in a study to classify the simulated pilot workload levels using a neural network based on BPs. A driver's mental overload is considered to be one of the most important contributors to traffic accidents. This article presents the first results of a comparison of the contributions of two categories of candidates (ERPs and BPs) for driver mental workload evaluation. ERPs and BPs were extracted from the recorded EEG data in a simulated driving task, the Lane Change Task and a combination task of the LCT and the PASAT. A very powerful classifier, Adaboost, was used in the present study. The comparison of these two kinds of data showed that the Band Powers could provide higher accuracy for driver mental workload classification.


    Access

    Access via TIB

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A comparison of classification for driver mental workload using ERP and band power parameters


    Additional title:

    Eine vergleichende Klassifizierung der mentalen Fahrerbelastung bei Anwendung von ERP und Bandleistungsparametern


    Contributors:


    Publication date :

    2009


    Size :

    4 Seiten, 1 Bild, 12 Quellen




    Type of media :

    Conference paper


    Type of material :

    Storage medium


    Language :

    English




    Driver Reliability Assessment Based Mental Workload

    Tang, Zhihui / Zhang, Kairan / Cheng, Xueqing | ASCE | 2009


    Driver Reliability Assessment Based Mental Workload

    Tang, Z. / Zhang, K. / Cheng, X. et al. | British Library Conference Proceedings | 2009


    Study of drivers mental workload model

    Miyashita, Y. / Miki, Y. / Itoh, K. et al. | British Library Conference Proceedings | 2009


    UAV Operator mental workload - A neurophysiological comparison of mental workload and vigilance

    Richards, Dale / Izzetoglu, Kurtulus / Shelton-Rayner, Graham | AIAA | 2017


    Automotive automation: effects, problems and implications for driver mental workload

    Young, M. S. / Stanton, N. A. / Cranfield University; College of Aeronautics; Department of Applied Psychology | British Library Conference Proceedings | 1997