Electrophysiological Degrading Correlates For Driving Attention Loss Threshold Determination
Statistics by Malaysian Institute of Road Safety Research (MIROS) showed that attention loss significantly lead to road accidents. Hence, the area of research on attention detection for driver safety is becoming more important. There have been a number of studies that displayed the possibility of id...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/24512/1/Electrophysiological%20Degrading%20Correlates%20For%20Driving%20Attention%20Loss%20Threshold%20Determination.pdf http://eprints.utem.edu.my/id/eprint/24512/2/Electrophysiological%20Degrading%20Correlates%20For%20Driving%20Attention%20Loss%20Threshold%20Determination.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Statistics by Malaysian Institute of Road Safety Research (MIROS) showed that attention loss significantly lead to road accidents. Hence, the area of research on attention detection for driver safety is becoming more important. There have been a number of studies that displayed the possibility of identifying drivers’ attention using electroencephalography (EEG) signal. The studies obtained the Event Related Potential (ERP) waveform from a small pool of samples. However, the data obtained were insufficient to significantly characterize attentiveness and inattentiveness due to the unique characteristic of each individual. Therefore, the aim of this research is to define the attentiveness state of each subject from large number of samples in controlled parameters to minimize the variability gap of the ERP peak between each individual. The experiment has been conducted using driving simulator to obtain the EEG data from two groups of subjects which were categorized as attentive and inattentive state by using two distinct stimulations i.e., listening to radio and no stimulation. The obtained results show significant boundary and similarity patterns for the level of attentiveness in both groups. Due to these patterns, a hybrid mean-fuzzy (HMF) technique was proposed to analyze the peak of N170 ERP decrement value versus the accident score based on the driving performance and attention threshold was determined accordingly. Three levels of attention namely ‘attentive’, ‘the beginning of inattentiveness’ and ‘inattentive’ state were presented within a new framework scale in the form of a fish bone diagram known as Attention Degradation Scale (ADS). In order to validate the feasibility of the proposed ADS for both groups, the analysis of the data has been done with and without ADS. Based on the outcome, 52% of the subjects were detected as attentive whilst 56% were in inattentive state which is significant as the percentage obtained with ADS was more than without it. Finally, a prototype application has been implemented to prove the theoretical data of attention level prediction. The prototype has successfully warned the subjects of potential accidents whenever the attention level was below the threshold value. Therefore, the findings of this research can be a promising foundation for alarm system which based on attention recognition technique that potentially would be able to reduce road accidents specifically with the proposed ADS. |
|---|