Publicação
Detection of drowsy driving using wearable sensors
| Resumo: | Drowsy driving is one of the leading causes of traffic accidents. Some solution provides feedback when the driver is drowsy, however, few tackle the issue in a way that allows for portability and early prevision. This study focuses on drowsiness detection during driving. Wearable sensors are used, for a low-cost, portable, automated, and non-intrusive solution. The wearable sensors chosen for biosignal acquisition are Empatica's E4 wristband for heart activity acquisition and Brainlink Pro for brain activity. Features were mainly in the time domain and time-frequency, and algorithms, such as Nearest Neighbours, Radial Basis Function, Support Vector Machine, Decision Tree, Random Forest, Multi-layer Perceptron, Naive Bayes, and Logistic Regression were trained and validated through the use of a database developed for this study (11 adults with normal last-night sleep, and 2 without any last-night sleep). Participants answered Pittsburgh, and Satisfaction, Alertness, Timing, Efficiency and Duration questionnaires, after which photoplethysmography and electroencephalography physiological signals were acquired during driving in a simulation environment. The practice-run discrimination and individual classification had comparable results, both slightly above average (70 to 80%). The evaluation metric values showed that the discrimination of sleep-deprived exams yielded significantly better. This suggests that the proposed methodology is capable of classifying sleep deprivation and surpasses existing ones in its portability |
|---|---|
| Autores principais: | Pereira, Duarte |
| Outros Autores: | Faria, Brígida Mónica; Reis, Luís Paulo; Faria, Brígida Mónica |
| Assunto: | Sleep prevention Driving simulation Biosignal acquisition Signal processing |
| Ano: | 2023 |
| País: | Portugal |
| Tipo de documento: | documento de conferência |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Instituto Politécnico do Porto |
| Idioma: | inglês |
| Origem: | Repositório Científico do Instituto Politécnico do Porto |
| Resumo: | Drowsy driving is one of the leading causes of traffic accidents. Some solution provides feedback when the driver is drowsy, however, few tackle the issue in a way that allows for portability and early prevision. This study focuses on drowsiness detection during driving. Wearable sensors are used, for a low-cost, portable, automated, and non-intrusive solution. The wearable sensors chosen for biosignal acquisition are Empatica's E4 wristband for heart activity acquisition and Brainlink Pro for brain activity. Features were mainly in the time domain and time-frequency, and algorithms, such as Nearest Neighbours, Radial Basis Function, Support Vector Machine, Decision Tree, Random Forest, Multi-layer Perceptron, Naive Bayes, and Logistic Regression were trained and validated through the use of a database developed for this study (11 adults with normal last-night sleep, and 2 without any last-night sleep). Participants answered Pittsburgh, and Satisfaction, Alertness, Timing, Efficiency and Duration questionnaires, after which photoplethysmography and electroencephalography physiological signals were acquired during driving in a simulation environment. The practice-run discrimination and individual classification had comparable results, both slightly above average (70 to 80%). The evaluation metric values showed that the discrimination of sleep-deprived exams yielded significantly better. This suggests that the proposed methodology is capable of classifying sleep deprivation and surpasses existing ones in its portability |
|---|
Atividades financiadas
Não foram encontrados projetos financiados.