Date of Award
Doctor of Philosophy
Civil, Construction, and Environmental Engineering
Shauna L. Hallmark
Research studies have shown work zones and rural intersections vulnerable to numerous crashes. In 2015 alone, there were around 96,700 crashes in work zones, an approximate 7.8% increase from 2014. Out of all the crashes, around 0.7% of the crashes involved at least one fatality with statistics showing work zone crashes occurring once every 5.4 minutes during that year (Facts and Statistics – Work Zone Safety, 2017). In addition, rural intersection crashes account for around 30% of crashes in rural areas with more than 80% of rural intersections fatalities occurring at rural unsignalized intersections (Golembiewski and Chandler, 2011). Crashes in rural areas are often severe because of higher approach speeds and longer emergency response times (Gonzales et al 2009). Past studies have given more priority to assess the safety effectiveness of various countermeasures mostly in terms of crash analysis both in work zones and rural intersections. However, little is known on the driving behavior of vehicles at advance warning area of work zones and driving behavior of vehicles at nonstop controlled approaches of rural intersections.
This study utilized SHRP 2 Naturalistic Driving Study (NDS) data and Roadway Information Database (RID). Using both the data set, the study developed statistical models to analyze driving behavior upstream of work zones and rural intersections.
The first study developed a mixed effect logistic regression model to analyze the driving behavior in advance warning area of work zones to find the effectiveness of different work zone signs. The result showed first work zone sign was not significantly affecting the driving behavior. Only speed limit, lane ends and CMS were found to be affecting the driving behavior. Active CMS was found to be more effective compared to not active CMS sign. Effect of overlapping signs was not found to have significant effect on the driving behavior. Speed limit with both work zone and feedback type were found to be significantly effective compared to normal speed limit signs with no indication of work zone. Speeding drivers were more likely to show response at different work zone signs with exception for drivers speeding at first sign. Distracted drivers were less likely to show response at work zone signs.
The second study built a mixed effect linear regression model to find different factors behind the response point of turning major street vehicles. The result showed that right turning vehicles started to show reaction to the turning maneuver slightly ahead to left turning vehicles. More than 70% of drivers showed reaction within 300 meters upstream of intersection for both types of turning maneuver. In addition, the study found driving speed at reaction point significantly affecting its location from intersection. Drivers speeding than the posted speed limit were associated with reaction point farther from the intersection.
In third study, a mixed effect logistic regression model was developed to find different factors affecting driving behavior of through moving vehicles at rural intersections. The result from this study showed that about 32% of drivers showed response to the intersections by decreasing speed by at least 3 miles per hour. Vehicles were more likely to show response to intersection at the time of presence of vehicles at the minor approaches. Non experienced drivers were found to be aware of the intersection ahead compared to experienced drivers. Drivers operating speed above 5 miles per hour were more likely to show response point. Intersections with intersection ahead warning signs was found to affect the response point positively.
Thapa, Raju, "Driving behavior analysis at work zones and rural intersections using SHRP 2 naturalistic driving data" (2019). Graduate Theses and Dissertations. 17332.