Characterization and Identification of Distraction During Naturalistic Driving Using Wearable Non-Intrusive Physiological Measure of Galvanic Skin Responses

Skin conductance;Galvanic skin responses;Driver distraction;Driver inattention;Distraction detection;GSR distraction detection;SVM-RFE;GSR feature analysis;Computer science;Computer and Information Science, College of Engineering & Computer Science

Rajendra, VikasBacha, Anys ;

University of Michigan

关键词: Skin conductance;    Galvanic skin responses;    Driver distraction;    Driver inattention;    Distraction detection;    GSR distraction detection;    SVM-RFE;    GSR feature analysis;    Computer science;    Computer and Information Science, College of Engineering & Computer Science;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/143521/Vikas%20Final%20Text%20Embedded.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Fatalities due to road accidents are mainly caused by distracted driving. Driving demands continuous attention of the driver. Certain levels of distraction while driving can cause the driver to lose his/her attention which might lead to a fatal accident. Thus, early detection of distraction will help reduce the number of accidents. Several researches have been conducted for automatic detection of driver distraction. Many previous approaches have employed camera-based techniques. However, these methods might detect the distraction rather late to warn the drivers. Although neurophysiological signals using Electroencephalography (EEG) have shown to be another reliable indicator of distraction,EEG signals are very complex, and the technology is intrusive to the drivers, which creates serious doubt for its implementation. In this thesis we investigate a non-intrusive physiological measure-Galvanic Skin Responses (GSR) using a wrist band wearable andconduct an empirical characterization of driver GSR signals during a naturalistic drivingexperiment. The proposed method is used to evaluate and extract statistical, frequency and time domain features to identify distraction. Also, several data mining techniques such as feature selection, feature-ranking, dimensionality reduction and feature space analysis are performed to generate discriminative bases that reduce the computational complexity for efficient identification of distraction using supervised learning. A signal processing technique: continuous decomposition analysis, exclusive for skin conductance signal was investigated to better understand the behavior of raw signal during cognitive and visual over load from secondary tasks while driving. The proposed driver monitoring and identification system on the edge provided evident results using GSR as a reliableindicator of driver distraction while meeting the requirement of early notification of distraction state to driver.