논문 상세보기
pp.83-92
Drowsy driving on highways is a critical traffic safety issue because the monotonous driving environment frequently induces driver fatigue, often leading to severe or fatal crashes. This study aims to develop a quantitative drowsy driving index (DDI) to proactively identify highrisk segments and establish effective safety management strategies. To achieve this, diverse static and dynamic traffic data along with weather conditions were integrated. We developed two distinct frequency models: a crash risk model setting drowsy driving crashes as the dependent variable, and a drowsiness-inducing risk model based on the frequency of drowsy driving behaviors. Poisson and negative binomial regression models were employed to account for the rare-event nature and overdispersion of crash data. Variables were carefully selected based on the variance inflation factor to prevent multicollinearity, and optimal models were determined using the Akaike information criterion. The final DDI was formulated by combining the quantified crash and drowsiness-inducing risks through a weight-based integration method. The analysis revealed a strong positive correlation between the proposed DDI and both actual drowsy driving crashes and behavioral frequencies. The proposed DDI framework offers a robust tool for monitoring and managing drowsy driving risks. Through public-private partnerships, this methodology can be utilized to provide real-time, location-based warnings and rest area guidance to drivers to significantly mitigate crash risks. Future research should expand the analysis to encompass a wider highway network and incorporate road geometry data, such as curve radii and gradients, to further enhance the precision and realism of the index.
1. 서론
2. 선행연구
2.1. 졸음운전 사고 주요 영향 요인 도출 연구
2.2. 졸음운전 행태 연구
3. 연구방법론
3.1. 데이터 수집 및 가공
3.2. 졸음운전 사고 및 행태 빈도 모형 개발
3.3. 다기준의사결정기법 기반의 DDI 개발
4. 연구 결과
4.1. 상관분석 결과
4.2. 모형 개발 결과
4.3. 졸음지수(DDI) 개발 결과
5. 결론
감사의 글
REFERENCES
