AbstractObjective: The objective of this research is to study the feasibility of measuring behavioral indicators that reflect effects of infrastructure and interaction with other road users.Methods: An observation study was performed using 6 cameras above a separated cycle path next to a road which included a crossing with both cyclists and cars. A learning method based on Single Shot MultiBox Detector was applied to automatically detect the cyclists, and cyclist tracks were determined. Next, kinematic parameters were calculated from the cyclists’ tracks. Amongst others, the cyclists’ intensity, speed, position on the cycle path, and the distance to each other were analyzed for a busy period as well as for a quiet period of the day.Results: With the measurement method developed in this study it is possible to analyze the cyclists’ intensity, the space they use at the cycle path, their average velocity, waiting times, the space and velocity amongst each other, and red light negation. However, collisions were not seen in the dataset analyzed, and the data is not sufficiently accurate to analyze sudden braking actions.Conclusion: It can be concluded that the developed measurement method provides insight of the cyclists’ behavior in such a way that it can already be used for obtaining information to make changes to the infrastructure that will improve the comfort and safety of cyclists. The method could be further developed for doing qualitative comfort and safety analyses, and for doing analyses of the interaction between various types of road participants. 相似文献
Objective: It is well known that alcohol and drugs influence driving behavior by affecting the central nervous system, awareness, vision, and perception/reaction times, but the resulting effect on driver injuries in car crashes is not fully understood. The purpose of this study was to identify factors affecting the injury severities of unimpaired, alcohol-impaired, and drug-impaired drivers.
Method: The current article applies a random parameters logit model to study the differences in injury severities among unimpaired, alcohol-impaired, and drug-impaired drivers. Using data from single-vehicle crashes in Cook County, Illinois, over a 9-year period from January 1, 2004, to December 31, 2012, separate models for unimpaired, alcohol-impaired, and drug-impaired drivers were estimated. A wide range of variables potentially affecting driver injury severity was considered, including roadway and environmental conditions, driver attributes, time and location of the crash, and crash-specific factors.
Results: The estimation results show significant differences in the determinants of driver injury severities across groups of unimpaired, alcohol-impaired, and drug-impaired drivers. The findings also show that unimpaired drivers are understandably more responsive to variations in lighting, adverse weather, and road conditions, but these drivers also tend to have much more heterogeneity in their behavioral responses to these conditions, relative to impaired drivers. In addition, age and gender were found to be important determinants of injury severity, but the effects varied significantly across all drivers, particularly among alcohol-impaired drivers.
Conclusions: The model estimation results show that statistically significant differences exist in driver injury severities among the unimpaired, alcohol-impaired, and drug-impaired driver groups considered. Specifically, we find that unimpaired drivers tend to have more heterogeneity in their injury outcomes in the presence potentially adverse weather and road surface conditions. This makes sense because one would expect unimpaired drivers to apply their full knowledge/judgment range to deal with these conditions, and the variability of this range across the driver population (with different driving experiences, etc.) should be great. In contrast, we find, for the most part, that alcohol-impaired and drug-impaired drivers have far less heterogeneity in the factors that affect injury severity, suggesting an equalizing effect resulting from the decision-impairing substance. 相似文献