Statistical analysis of weaving before and after managed lane conversion
Araque Rojas, Santiago Andres
MetadataShow full item record
This thesis presents a statistical analysis of weaving in a managed lane system which is evolving from a High-Occupancy Vehicle (HOV) system to a High-Occupancy Toll system (HOT). Weaving was, assessed along the I-85 corridor in Atlanta, during three different phases in the conversion from HOV to HOT: 1) the existing HOV managed lane system prior to conversion to HOT lanes, 2) after restriping of some weaving zones but prior to conversion of the HOV lanes to HOT lanes and, 3) after the HOT managed lane system opened. Each phase was analyzed to see how weaving behavior into and out of the managed lane system was affected by changes in the system. To accomplish the analysis, video was collected using Georgia Department of Transportation cameras along the corridor. The videos were transferred to an Android Tablet, in which an App developed by the research team was used to record data from the videos. Using the processed weaving data, a comparison of weaving activity during each phase was performed. Data were also analyzed across time of day, speed differentials, and whether the weaves in question were performed legally (within established weaving zones) or illegally (across double-solid striped lane markings). After a comparison of weaving behavior along different variables, a regression tree analysis was completed. The analysis showed that weaving intensity increased as the system was converted from HOV to HOT. However, illegal weaving decreased significantly once the HOT system was in place, perhaps due to stricter enforcement or perhaps due to driver response to illegally entering and leaving tolled lanes. The regression tree analyses indicated that weaving intensity was highly dependent upon whether it was legal or illegal to weave and upon the phase of conversion during which the weave occurred.