|M.Sc Student||Zoher Elias|
|Subject||A Model for Traffic Event Prediction in Freeways|
|Department||Department of Civil and Environmental Engineering||Supervisor||Professor Toledo Tomer|
|Full Thesis text - in Hebrew|
The aim of this study was to examine in depth the influence of traffic characteristics and time characteristics on incident occurrence in highways. The traffic incidents that are studied included crashes with injured passengers, or only with property damages to the vehicle, objects on road, stalled (stuck) cars, damages in road infrastructure, natural disturbances and so on. These are random incidents that occur relatively frequently.
The main purpose of this study is to examine if there is a statistically significant relationship between the rate of these traffic incidents and traffic flow characteristics and to study the effect of these characteristics on the probability of traffic incidents occurrence. For this purpose statistical analysis of data and development of a disaggregate event occurrence model were carried out.
For the purpose of this study, data from 18 sections of the Netivey Ayalon highway in Tel Aviv were examined. These sections cover a total distance of approximately 29 kilometers. Data on traffic flow characteristics was collected by magnetic loop detectors during three years, from 2004 to 2006. In addition, records of traffic incidents that occurred in the same highway sections during the same period of time were used. The data shows that on average approximately 12 traffic incidents occur daily on this road.
The results of the model estimation showed that there is a statistically significant relationship between the probability of traffic incident occurrence and the traffic flow and time characteristics. Traffic characteristics that influence this probability are speed and flow. They both exhibit a U-shaped relation to incident occurrence, with lower rates in low and high speeds, and in low and high flows. The relationship between the time of day variables and the probability to incident occurrence is also statistically significant. After accounting for differences in traffic characteristics, the probability of incident occurrence is higher at night than during the day time.