Syed Mir Talha Zobaed | University of Rajshahi, Bangladesh (original) (raw)
Uploads
Papers by Syed Mir Talha Zobaed
In recent years, driver‘s drowsiness has been one of the major causesof mortality in road acciden... more In recent years, driver‘s drowsiness has been one of the major causesof mortality in road accidents worldwide and can lead to severe physical injuries,deaths and significant and noticeable economic losses. Many of these road accidentsand deaths could be avoided, if driver‘s drowsiness could be properly monitored anddrivers are given early warnings. In this thesis work, novel but simplemethods has been proposed to detect driver‘s drowsiness or sleeponset using single channel EEG signal analysis. Block sample entropy, SDSis, median absolute deviation and summation of median absolute deviation (SumMAD) are extracted as features. Threshold has been calculated from the wake signal and this threshold is used for classification of unknown epochs. The superiority of this work is to identify the sleep onset by using only one feature effective for practical implementation. It does not require any training and hence, new subject adaptation is comparatively easier and real time implementation. The proposed approach can be easily implemented in smart device for drowsiness detection and alarming system for vehicle‘s driver. The publicly available dataset is used to evaluate the performance of the proposed method. The experimental results show that newly introduced approach performs better than that of the state-of-the-art algorithms.
In recent years, driver‘s drowsiness has been one of the major causesof mortality in road acciden... more In recent years, driver‘s drowsiness has been one of the major causesof mortality in road accidents worldwide and can lead to severe physical injuries,deaths and significant and noticeable economic losses. Many of these road accidentsand deaths could be avoided, if driver‘s drowsiness could be properly monitored anddrivers are given early warnings. In this thesis work, novel but simplemethods has been proposed to detect driver‘s drowsiness or sleeponset using single channel EEG signal analysis. Block sample entropy, SDSis, median absolute deviation and summation of median absolute deviation (SumMAD) are extracted as features. Threshold has been calculated from the wake signal and this threshold is used for classification of unknown epochs. The superiority of this work is to identify the sleep onset by using only one feature effective for practical implementation. It does not require any training and hence, new subject adaptation is comparatively easier and real time implementation. The proposed approach can be easily implemented in smart device for drowsiness detection and alarming system for vehicle‘s driver. The publicly available dataset is used to evaluate the performance of the proposed method. The experimental results show that newly introduced approach performs better than that of the state-of-the-art algorithms.