교통사고와 속도 패턴 모형 -국도36호선 증평-충주구간을 중심으로 - Models of Traffic Accident and Speed Pattern In the Case of Jeungpyeong-Chungju Section, National Road No.36원문보기
This study deals with the relations between the speed pattern and traffic accidents. The goal is to develop the models of their relations. In pursuing the above, this study gives particular attentions to collecting the data of speed pattern, geometric structure and accident. The data of speed patter...
This study deals with the relations between the speed pattern and traffic accidents. The goal is to develop the models of their relations. In pursuing the above, this study gives particular attentions to collecting the data of speed pattern, geometric structure and accident. The data of speed pattern was collected from the 18 Automatic Vehicle Number Identification System (AVI) installed on 47.1km section from Hwasung-ri, Jeungpyong to Dalcheon-dong, Chungju. The accident data were retrieved from TAAS system of Korea Road Traffic authority(KoROAD). Otherwise, the data of geometric structure, traffic volume, and others were collected from the field survey. The main results analyzed are as follow. First, 4 null hypothesis tests were all rejected. It was analyzed that the speed patterns were not appropriate to the independent variables of accident model. Second, 3 multi-linear, 3 multi-nonlinear and 3 Poisson regression models which were statistically significant(with values above 0.856 and values above 0.299) except 3 speed variance models were developed. In these models, 4 dependent variables(the number of accident, EPDO, average speed and speed variance) and 23 independent variables were utilized. Finally, the tests of RMSE(root mean square error), %RMSE(root mean square error percent), MPB(mean prediction bias) and MAD(mean absolute deviation), showed that the multiple nonlinear regression model in the number of traffic accidents, EPDO and speed variance, and Poisson regression model in the average speed were evaluated to be well fitted to the given data.
This study deals with the relations between the speed pattern and traffic accidents. The goal is to develop the models of their relations. In pursuing the above, this study gives particular attentions to collecting the data of speed pattern, geometric structure and accident. The data of speed pattern was collected from the 18 Automatic Vehicle Number Identification System (AVI) installed on 47.1km section from Hwasung-ri, Jeungpyong to Dalcheon-dong, Chungju. The accident data were retrieved from TAAS system of Korea Road Traffic authority(KoROAD). Otherwise, the data of geometric structure, traffic volume, and others were collected from the field survey. The main results analyzed are as follow. First, 4 null hypothesis tests were all rejected. It was analyzed that the speed patterns were not appropriate to the independent variables of accident model. Second, 3 multi-linear, 3 multi-nonlinear and 3 Poisson regression models which were statistically significant(with values above 0.856 and values above 0.299) except 3 speed variance models were developed. In these models, 4 dependent variables(the number of accident, EPDO, average speed and speed variance) and 23 independent variables were utilized. Finally, the tests of RMSE(root mean square error), %RMSE(root mean square error percent), MPB(mean prediction bias) and MAD(mean absolute deviation), showed that the multiple nonlinear regression model in the number of traffic accidents, EPDO and speed variance, and Poisson regression model in the average speed were evaluated to be well fitted to the given data.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.