This study aims to compare three distinct methods used to compute access density and provide a comprehensive weighted methodology to enable standardization for research and application in the future. Access density is a widely used concept that calculates the number of access points within a given distance and has been extensively applied to studies related to crash modeling, operational impact and planning. Methods used in past research show that access density is computed differently by different studies and all studies do not include all access points. The weighted methodology proposed takes into account all access points including driveways, intersections and median openings and categorizes them into geometric combinations. Each geometric combination have potential number of conflict points which include diverging, weaving, merging and crossing movements depending on the type of access point. Weights were assigned to each geometry type based on these conflict point ratio. In conclusion the study identifies and compares methods previously used to compute access density and accordingly, recommends a weighted methodology that includes all access points which can be used as a standard, universal measure all access density related studies including but not limited to safety impacts, operational impacts and planning guidelines.

Many studies have shown that intersections are among the most dangerous locations of a roadway network. Therefore, there is a need to understand the factors that contribute to traffic crashes at such locations. One approach is to model crash occurrences based on configuration, geometric characteristics and traffic. Instead of combining all variables and crash types to create a single statistical model, this analysis created several models that address the different factors that affect crashes, by type of collision as well as injury level, at signalized intersections. The first objective was to determine if there is a difference between important variables for models based on individual crash types or severity levels and aggregated models. The second objective of this research was to investigate the quality and completeness of the crash data and the effect that incomplete data has on the final results. A detailed and thorough data collection effort was necessary for this research to ensure the quality and completeness of this data. Multiple agencies were contacted and databases were crosschecked (i.e. state and local jurisdictions/agencies). Information (including geometry, configuration and traffic characteristics) was collected for a total of 832 intersections and over 33,500 crashes from Brevard, Hillsborough and Seminole Counties and the City of Orlando. Due to the abundance of data collected, a portion was used as a validation set for the tree-based regression. Hierarchical tree-based regression (HTBR) and ordered probit models were used in the analyses. HTBR was used to create models for the expected number of crashes for collision type as well as injury level. Ordered probit models were only used to predict crash severity levels due to the ordinal nature of this dependent variable. Finally, both types of models were used to predict the expected number of crashes. More specifically, tree-based regression was used to consider the difference in the relative importance of each variable between the different types of collisions. First, regressions were only based on crashes available from state agencies to make the results more comparable to other studies. The main finding was that the models created for angle and left turn crashes change the most compared to the model created from the total number of crashes reported on long forms (restricted data usually available at state agencies). This result shows that aggregating the different crash types by only estimating models based on the total number of crashes will not predict the number of expected crashes as accurately as models based on each type of crash separately. Then, complete datasets (full dataset based on crash reports collected from multiple sources) were used to calibrate the models. There was consistently a difference between models based on the restricted and complete datasets. The results in this section show that it is important to include minor crashes (usually reported on short forms and ignored) in the dataset when modeling the number of angle or head-on crashes and less important to include minor crashes when modeling rear-end, right turn or sideswipe crashes. This research presents in detail the significant geometric and traffic characteristics that affect each type of collision. Ordered probit models were used to estimate crash injury severity levels for three different types of models; the first one based on collision type, the second one based on intersection characteristics and the last one based on a significant combination of factors in both models. Both the restricted and complete datasets were used to create the first two model types and the output was compared. It was determined that the models based on the complete dataset were more accurate. However, when compared to the tree-based regression results, the ordered probit model did not predict as well for the restricted dataset based on intersection characteristics. The final ordered probit model showed that crashes involving a pedestrian/bicyclist have the highest probability of a severe injury. For motor vehicle crashes, left turn, angle, head-on and rear-end crashes cause higher injury severity levels. Division (a median) on the minor road, as well as a higher speed limit on the minor road, was found to lower the expected injury level. This research has shed light on several important topics in crash modeling. First of all, this research demonstrated that variables found to be significant in aggregated crash models may not be the same as the significant variables found in models based on specific crash types. Furthermore, variables found to be significant in crash type models typically changed when minor crashes were added to complete the dataset. Thirdly, ordered probit models based on significant crash-type and intersection characteristic variables have greater crash severity prediction power, especially when based on the complete dataset. Lastly, upon comparison between tree-based regression and ordered probit models, it was found that the tree-based regression models better predicted the crash severity levels.

Statistics show that signalized intersections are among the most dangerous locations of a roadway network. Different approaches including crash frequency and severity models have been used to establish the relationship between crash occurrence and intersection characteristics. In order to model crash occurrence at signalized intersections more efficiently and eventually to better identify the significant factors contributing to crashes, this dissertation investigated the temporal, spatial, and site correlations for total, rear-end, right-angle and left-turn crashes. Using the basic regression model for correlated crash data leads to invalid statistical inference, due to incorrect test statistics and standard errors based on the misspecified variance. In this dissertation, the Generalized Estimating Equations (GEEs) were applied, which provide an extension of generalized linear models to the analysis of longitudinal or clustered data. A series of frequency models are presented by using the GEE with a Negative Binomial as the link function. The GEE models for the crash frequency per year (using four correlation structures) were fitted for longitudinal data; the GEE models for the crash frequency per intersection (using three correlation structures) were fitted for the signalized intersections along corridors; the GEE models were applied for the rear-end crash data with temporal or spatial correlation separately. For right-angle crash frequency, models at intersection, roadway, and approach levels were fitted and the roadway and approach level models were estimated by using the GEE to account for the "site correlation"; and for left-turn crashes, the approach level crash frequencies were modeled by using the GEE with a Negative Binomial link function for most patterns and using a binomial logit link function for the pattern having a higher proportion of zeros and ones in crash frequencies. All intersection geometry design features, traffic control and operational features, traffic flows, and crashes were obtained for selected intersections. Massive data collection work has been done. The autoregression structure is found to be the most appropriate correlation structure for both intersection temporal and spatial analyses, which indicates that the correlation between the multiple observations for a certain intersection will decrease as the time-gap increase and for spatially correlated signalized intersections along corridors the correlation between intersections decreases as spacing increases. The unstructured correlation structure was applied for roadway and approach level right-angle crashes and also for different patterns of left-turn crashes at the approach level. Usually two approaches at the same roadway have a higher correlation. At signalized intersections, differences exist in traffic volumes, site geometry, and signal operations, as well as safety performance on various approaches of intersections. Therefore, modeling the total number of left-turn crashes at intersections may obscure the real relationship between the crash causes and their effects. The dissertation modeled crashes at different levels. Particularly, intersection, roadway, and approach level models were compared for right-angle crashes, and different crash assignment criteria of "at-fault driver" or "near-side" were applied for disaggregated models. It shows that for the roadway and approach level models, the "near-side" models outperformed the "at-fault driver" models. Variables in traffic characteristics, geometric design features, traffic control and operational features, corridor level factor, and location type have been identified to be significant in crash occurrence. In specific, the safety relationship between crash occurrence and traffic volume has been investigated extensively at different studies. It has been found that the logarithm of traffic volumes per lane for the entire intersection is the best functional form for the total crashes in both the temporal and spatial analyses. The studies of right-angle and left-turn crashes confirm the assumption that the frequency of collisions is related to the traffic flows to which the colliding vehicles belong and not to the sum of the entering flows; the logarithm of the product of conflicting flows is usually the most significant functional form in the model. This study found that the left-turn protection on the minor roadway will increase rear-end crash occurrence, while the left-turn protection on the major roadway will reduce rear-end crashes. In addition, left-turn protection reduces Pattern 5 left-turn crashes (left-turning traffic collides with on-coming through traffic) specifically, but it increases Pattern 8 left-turn crashes (left-turning traffic collides with near-side crossing through traffic), and it has no significant effect on other patterns of left-turn crashes. This dissertation also investigated some other factors which have not been considered before. The safety effectiveness of many variables identified in this dissertation is consistent with previous studies. Some variables have unexpected signs and a justification is provided. Injury severity also has been studied for Patterns 5 left-turn crashes. Crashes were located to the approach with left-turning vehicles. The "site correlation" among the crashes occurred at the same approach was considered since these crashes may have similar propensity in crash severity. Many methodologies and applications have been attempted in this dissertation. Therefore, the study has both theoretical and implementational contribution in safety analysis at signalized intersections.

Traffic conflicts associated to signalized intersections are one of the major contributing factors to crash occurrences. Driver behavior plays an important role in the safety concerns related to signalized intersections. In this research effort, dynamics of driver behavior in relation to the traffic conflicts occurring at the onset of yellow is investigated. The area ahead of intersections in which drivers encounter a dilemma to pass through or stop when the yellow light commences is called Dilemma Zone (DZ). Several DZ-protection algorithms and advance signal settings have been developed to accommodate the DZ-related safety concerns. The focus of this study is on drivers' decision dynamics, human learning, and choice behavior in DZ, and DZ-related safety measures. First, influential factors to drivers' decision in DZ were determined using a driver behavior survey. This information was applied to design an adaptive experiment in a driving simulator study. Scenarios in the experimental design are aimed at capturing drivers learning process while experiencing safe and unsafe signal settings. The result of the experiment revealed that drivers do learn from some of their experience. However, this learning process led into a higher level of risk aversion behavior. Therefore, DZ-protection algorithms, independent of their approach, should not have any concerns regarding drivers learning effect on their protection procedure. Next, the possibility of predicting drivers' decision in different time frames using different datasets was examined. The results showed a promising prediction model if the data collection period is assumed 3 seconds after yellow. The prediction model serves advance signal protection algorithms to make more intelligent decisions. In the next step, a novel Surrogate Safety Number (SSN) was introduced based on the concept of time to collision. This measure is applicable to evaluate different DZ-protection algorithms regardless of their embedded methodology, and it has the potential to be used in developing new DZ-protection algorithms. Last, an agent-based human learning model was developed integrating machine learning and human learning techniques. An abstracted model of human memory and cognitive structure was used to model agent's behavior and learning. The model was applied to DZ decision making process, and agents were trained using the driver simulator data. The human learning model resulted in lower and faster-merging errors in mimicking drivers' behavior comparing to a pure machine learning technique. / Ph. D.

Among the most critical elements at signalized intersections are the design of vehicle detection equipment and the timing of change and clearance intervals. Improperly timed clearance intervals or improperly placed detection equipment can potentially place drivers in a Type I dilemma zone, where approaching motorists can neither proceed through the intersection before opposing traffic is released nor safely stop in advance of the stop bar. Type II dilemma zones are not necessarily tied to failures in design, but are more readily tied to difficulties in driver decision making associated with comprehension and behavior. The Type II dilemma zone issues become even more prevalent at high-speed intersections where there is greater potential for serious crashes and more variability in vehicle operating speeds. This research initiative attempts to further describe the impact of driver behavior and comprehension on dilemma zones. To address this notion several experiments are proposed. First, a large empirical observation of high-speed signalized intersections is undertaken at 10 intersection approaches in Vermont. This resulted in the collection of video and speed data as well as full intersection inventories and signal timings. These observations are reduced and analyzed for the purpose of reexamining the boundaries of a Type II dilemma zone. Second, a comparison of point and space sensors for the purpose of dilemma zone mitigation was conducted. This experiment provides evidence supporting the notion that space sensors have the potential for providing superior dilemma zone protection. Third, a computer based survey is conducted to identify if drivers comprehend the correct meaning of the solid yellow indication and how this relates to their predicted behavior. Lastly, a regression model is developed drawing on the data collected from the field observation as well as the static survey to determine how characteristics such as the speed and position of the vehicle as well as driver age and experience influence driver behavior in the Type II dilemma zone. Cumulatively, these experiments will shed additional light on the influence of driver behavior and comprehension on the Type II dilemma zone.

Crash data suggests that intersections are areas producing conflicts among the various road users because of entering and crossing movements. Traffic signal control systems may not always be sufficient in preventing collisions at intersections between emergency and other vehicles. The Firefighter Fatality Retrospective Study of 2002 illustrates that the second leading cause of fatal injury for firefighters is vehicle collisions. Furthermore, the involvement of an emergency vehicle in a crash can negatively affect the overall efficiency of emergency response services. Thus, there is a need to facilitate the implementation of higher-payoff strategies to improve the safety of emergency vehicle passage through signalized intersections. This research aims to provide a basis for the transportation professionals to identify problem areas and take measures that will potentially enhance intersection safety for emergency vehicles. It includes the presentation and comparison of the EV crash situation in Northern Virginia. The results indicate that 49% of all EV accidents along U.S. Highways in Northern Virginia occurred at signalized intersections. This percentage is 75% along U.S. Highways in Fairfax County, the largest county in Northern Virginia, and it is 79% along U.S. 1 in Fairfax County. The analysis, also, illustrates that the major collision type at signalized intersections was of the angle type, which suggests that an appropriate warning sign may be absent. These findings enhance our understanding of emergency vehicle crash characteristics and thus, may facilitate the identification of possible warrants to be used in determining the appropriateness of installing signal preemption equipment at signalized intersections. / Master of Science

Crash prediction models are used to estimate the number of crashes using a set of explanatory variables. The highway safety community has used modeling techniques to predict vehicle-to-vehicle crashes for decades. Specifically, generalized linear models (GLMs) are commonly used because they can model non-linear count data such as motor vehicle crashes. Regression models such as the Poisson, Zero-inflated Poisson (ZIP), and the Negative Binomial are commonly used to model crashes. Until recently very little research has been conducted on crash prediction modeling for pedestrian-motor vehicle crashes. This thesis considers several candidate crash prediction models using a variety of explanatory variables and regression functions. The goal of this thesis is to develop a pedestrian crash prediction model to contribute to the field of pedestrian safety prediction research. Additionally, the thesis contributes to the work done by the Federal Highway Administration to estimate pedestrian exposure in urban areas. The results of the crash prediction analyses indicate the pedestrian-vehicle crash model is similar to models from previous work. An analysis of two pedestrian volume estimation methods indicates that using a scaling technique will produce volume estimates highly correlated to observed volumes. The ratio of crash and exposure estimates gives a crash rate estimation that is useful for traffic engineers and transportation policy makers to evaluate pedestrian safety at signalized intersections in an urban environment. / Master of Science

Wider bike facilities intuitively accommodate a greater number of cyclists in the same amount of time, but specific queue discharge characteristics associated with varying widths and/or types of bike facilities have not been thoroughly documented. The focus of this research analyzed queues of cyclists at four signalized intersections in Portland, OR with varying widths on the approach and downstream intersection legs. A total of 2,820 cyclists within 630 groups of queued cyclists were observed at five different intersection layouts in Portland, Oregon. The layouts consisted of: a standard bike lane six feet wide connecting bicyclists to a standard bike lane six feet wide, a standard bike lane five feet wide connecting bicyclists to two standard bike lanes each five feet wide, a buffered bike lane 12 feet wide connecting bicyclists to a standard bike lane 6.5 feet wide, a bike box 21 feet wide connecting bicyclists to a buffered bike lane 10 feet wide, and a bike box 15 feet wide connecting bicyclists to two standard bike lanes each five feet wide. For each configuration, the following aspects were analyzed: average headway per cyclist within each queue, the time required for queues to enter the intersection, the time required for queues to clear the intersection, the number of cyclists within queues, the width of the bicycle facilities, the approach grade, and the utilization of a bike box at the intersection approach if it was present. The first major focus of the analysis reviewed the average headway values associated with each observed queue of cyclists. The queue size with the lowest mean of the average headway was for groups of seven cyclists with an average headway of approximately 0.8 seconds per cyclist. For queues larger than seven in size, the mean of the average headway remained relatively stable until queues of 12 in size and started to slightly increase toward approximately 1.0 seconds for queues larger than 12 cyclists. In addition, it appears that utilization of a bike box has a potential relationship with a reduced average headway as compared to queues that do not utilize a bike box. The associated reduction in the mean of the average headway was approximately 0.2 to 0.3 seconds per cyclist for queues of three or more in size. The second major focus of the analysis reviewed the queue discharge rate associated with each observed queue of cyclists. The results appear to potentially indicate that wider bike facilities approaching an intersection, wider receiving bike facilities, or utilization of a bike box generally discharge queues of bicyclists into the intersection over a shorter amount of time as compared to facilities that are narrower or underutilized. The installation of a bike box at one of the study intersections increased the approach width from five to 15 feet and resulted in consistently lower average discharge times for all queue sizes, a reduction of greater than one second for queues of two cyclists to as much as about four seconds for queues of nine cyclists. The third major focus of the analysis reviewed the intersection clearance time associated with each observed queue of cyclists. The results appear to potentially indicate that wider bike facilities approaching an intersection, wider receiving bike facilities, or utilization of a bike box generally clear queues of bicyclists through the intersection over a shorter amount of time as compared to facilities that are narrower or underutilized.

The role of walking in the development of healthy, livable communities is being increasingly recognized. In urban areas, intersections represent locations where different modes converge, and are often viewed as deterrents to walking. This is due to the unwarranted and often unnecessary delays imposed by signal timing policies for pedestrians and increased potential for conflicts. Traditional signal timing design prioritizes vehicles over pedestrians leading to undesirable consequences such as large delays and risky pedestrian behaviors. Pedestrians are accommodated in a manner that is designed to cause least interruption to the flow of motor vehicles. This lack of pedestrian accommodation at signalized intersections is the focus of this dissertation. Understanding pedestrian attitudes and perceptions is important because it offers insights into actual crossing behavior at signalized intersections. An intercept survey of 367 crossing pedestrians was undertaken at four signalized intersections in Portland, Oregon, and binary logistic regression models were constructed to quantify the impacts of demographics, trip characteristics and type of infrastructure on pedestrian perceptions and attitudes regarding delay, crossing time and motivators for crossing decisions. Safety was found to have a larger effect than compliance on the decision to cross the street. Pedestrians at recall intersections expressed higher satisfaction with delay than at actuated intersections. Novel methods to measure pedestrian delay using 2070 signal controllers and Voyage software were developed. These methods have been adopted by the City of Portland to record actuation trends and delays at various intersections. In the absence of demand data, pedestrian push button actuations can be considered as a proxy for crossing demand. The micro-simulation software VISSIM was used to analyze delays resulting from varying pedestrian and vehicle volumes on a network of three intersections in Portland, Oregon. From a pedestrian perspective, free operation was found to be always beneficial due to lower pedestrian delays. However, from a system wide perspective, free operation was found to be beneficial only under low-medium traffic conditions from an overall delay reduction viewpoint, while coordinated operation showed benefits under heavy traffic conditions, irrespective of the volume of pedestrians. Control strategies were developed to identify the best mode of signal controller operation that produced the lowest overall average delay per user. A procedure to identify the optimal control strategy based on user inputs (major street volume to capacity ratios and rate of pedestrian phase serviced for the minor street) was developed. The procedure was applied to a network of three intersections in east Portland, OR and the findings were verified. This research offers significant contributions in the field of pedestrian research. The findings related to attitudes and perceptions of crossing pedestrians offer greater insights into pedestrian crossing behavior and add to the body of existing literature. The methods developed to obtain pedestrian actuations and delay data from signal controllers represent an easy and cost-effective way to characterize pedestrian service at intersections. The results pertaining to signal timing strategies represent an important step towards incorporating pedestrian needs at intersections and demonstrate how control strategies employed to benefit pedestrians could benefit the entire system.

Since its inception in 1965, the Level-of-Service (LOS) has proved to be an important and practical "quality of service" indicator for transportation facilities around the world, widely used in the transportation and planning fields. The LOS rates these facilities' traffic operating conditions through the following delay-based indicators (ordered from best to worst conditions): A, B, C, D, E and F. This LOS rating has its foundation on quantifiable measures of effectiveness (MOEs) and on road users' perceptions; altogether, these measures define a LOS based on acceptable traffic operating conditions for the road user, implying that traffic safety is inherent to this definition. However, since 1994 safety has been excluded from the LOS definition since it cannot be quantified nor explicitly defined. The latter has been the motivation for research based on the LOS-Safety relationship, conducted at the University of Central Florida (UCF). Using data from two of the most studied transportation facility types within the field of traffic safety, signalized intersections and multilane high-speed arterial corridors, the research conducted has the following main objectives: to incorporate the LOS as a parameter in several traffic safety models, to extend the methodology adopted in previous studies to the subject matter, and to provide a platform for future transportation-related research on the LOS-Safety relationship. A meticulous data collection and preparation process was performed for the two LOS-Safety studies comprising this research. Apart from signalized intersections' and multilane-high speed arterial corridors' data, the other required types of information corresponded to crashes and road features, both obtained from FDOT's respective databases. In addition, the Highway Capacity Software (HCS) and the ArcGIS software package were extensively used for the data preparation. The result was a representative and robust dataset for each LOS-Safety study, to be later tested and analyzed with appropriate statistical methods. Regarding the LOS-Safety study for signalized intersections, two statistical techniques were used. The Generalized Estimating Equations (GEEs), the first technique, was used for the analyses considering all periods of a regular weekday (i.e. Monday through Friday): Early Morning, A.M. Peak, Midday, P.M. Peak and Late Evening; the second technique considered was the Negative Binomial, which was used for performing an individual analysis per period of the day. On the other hand, the LOS-Safety study for multilane high-speed arterial corridors made exclusive use of the Negative Binomial technique. An appropriate variable selection process was required for the respective model building and calibration procedures; the resulting models were built upon the six following response variables: total crashes, severe crashes, as well as rear-end, sideswipe, head-on and angle plus left-turn crashes. The final results proved to be meaningful for the understanding of traffic congestion effects on road safety, and on how they could be useful within the transportation planning scope. Overall, it was found that the risk for crash occurrence at signalized intersections and multilane high-speed arterial corridors is quite high between stable and unacceptable operating conditions; it was also found that this risk increases as it becomes later in the day. Among the significant factors within the signalized intersection-related models were LOS for the intersection as a whole, cycle length, lighting conditions, land use, traffic volume (major and minor roads), left-turn traffic volume (major road only), posted speed limit (major and minor roads), total number of through lanes (major and minor roads), overall total and total number of left-turn lanes (major road only), as well as county and period of the day (dummy variables). For multilane-high speed arterial corridors, the final models included LOS for the road section, average daily traffic (ADT), total number of through lanes in a single direction, total length of the road section, pavement surface type, as well as median and inside shoulder widths. A summary of the overall results per study, model implications and each LOS indicator is presented. Some of the final recommendations are to develop models for other crash types, to perform a LOS-Safety analysis at the approach-level for signalized intersections, as well as one that incorporates intersections within the arterial corridors' framework.