Accident prediction by conflict study

Traffic conflict studies are used in diagnosis of safety and operational problems at a roadway location. The basic aim of the traffic conflict technique is to expand the range of safety research from accidents to potential accidents. Therefore, the objective of this work is to study and design accident prediction models using the serious conflicts, and traffic volume, to predict potential accidents at unsignalized intersections (T junctions). This study was carried out at ten unsignalized intersections (T junction) in the State of Selangor. Conflict and volume data was collected on weekdays. One day of data was collected for each site. Manual records supported by video method was used to capture the conflict data, and volume data over period of observation (from 7.00 to 13.00) and (from 15.00 to 18.00). Conflict and volume data were recorded within each hour of the study period. Regression analysis using EXCEL, and SPSS software was used to obtain significant correlate between the variables. As in the earlier study which found accident data and serious conflict data well correlated, in this study the result obtained indicates that serious conflict and accident figures are well correlated and the types of vehicles involved in the conflicts correlated well with the types of vehicles involved in the accidents. The relationship between accidents and traffic volume (veh/hr), and between conflicts and traffic volume (veh/hr), indicates that the number of traffic accidents and conflicts increases when the amount of traffic volume increases. The result obtained from the relationship between traffic accident and total traffic volume over nine hours indicates that traffic accidents and total traffic volume correlate well. And it suggests that the total traffic volume can be used to predict accident potential for any similar junctions with the same layout characteristics.

Text FK 2000 74 IR.pdf Download (13MB)

Actions (login required)

View Item