Mid-block crosswalks present a challenging detection task when drivers are not familiar with the area, crosswalks are at locations with high traffic volume, or the visibility of the approaching crosswalk is obstructed. A major contributor to crashes at uncontrolled marked mid-block crosswalks on multilane roads is the presence of a motorist who is yielding to a pedestrian in the crosswalk, creating an additional threat from a motorist in the adjacent lane who may not see the pedestrian. Advance yield markings and sign prompts have been shown to be effective when there is no obstruction in either travel lane. The purpose of this research is to evaluate drivers' performance in response to multi-threat (obstruction in travel lane) and sight-limited (obstruction in parking lane) scenarios with and without advance yield markings and sign prompts. No direct evaluations of the effectiveness of this countermeasure in these scenarios have been undertaken. The methods used to evaluate drivers' performance include observations of drivers navigating virtual crosswalks on a driving simulator, field observations of drivers responding to staged pedestrian crossings, and observations of drivers in an actual vehicle on the open roadway navigating marked mid-block crosswalks. Dependent variables include yielding behavior and eye-fixations. It is hypothesized that at multi-threat and sight-limited scenarios drivers approaching a crosswalk are more likely to look for pedestrians, to yield and to yield sooner when advance yield markings and prompt sign are used than when the traditional markings are used. The extent of these benefits is highly dependent on visibility improvements at the site, providing real benefit only when sight distance is adequate. This comprehensive evaluation, limited to low speed approaches in an urban environment, provides data from sight-limited and multi-threat scenarios where the probability of seeing the pedestrian is low. This information can help traffic engineers make informed decisions when implementing pedestrian safety countermeasures. It also provides data from locations in the United States to add to the studies completed in Canada which served as the basis for inclusion of the advance yield markings and related sign as guidance in the Manual on Uniform Traffic Control Devices.

Older drivers are over-represented in angled crashes when compared with younger experienced drivers. Past research primarily points to age-related cognitive and physical decline, which can impede older drivers' ability to monitor their driving environment efficiently and decrease their ability to maintain adequate situational awareness. Despite compensatory behaviors such as driving less, driving more slowly or avoiding driving in inclement conditions, there is evidence that in some cases these drivers may be under-compensating, as older drivers are still involved in more angled crashes than any other category. Of particular concern are intersections in which other vehicles can approach from the side. Two experiments described here investigate whether tailored feedback based on a driver's own unsafe behaviors and active, situation-based training in a simulator can change drivers' attitudes about their own abilities, raise their awareness of the crash risks for older drivers and lead to long-term improvements of driving behavior such as increased side-to-side scanning while negotiating intersections. Experiment 1 investigated whether customized feedback tailored to the individual's specific unsafe driving behaviors in a simulator can successfully alter an older driver's perceptions of his driving skills. Experiment 2 compared how effectively customized feedback about a driver's specific unsafe driving behaviors on the open road followed by active situation-based training in a simulator can improve road scanning and head turning behavior when compared with lecture-style training. The results from Experiment 1 demonstrated that letting drivers make errors in a simulator and then providing customized feedback was successful in changing older drivers' perception of their ability, making them more willing to change driving behavior. The results from Experiment 2 indicated that capturing drivers' errors on the road, providing customized feedback, and then adding active training in a simulator increased side-to-side scanning in intersections by nearly 100% in both post-training simulator and field drives. A second group, which received passive classroom-style training, demonstrated no significant improvement. In summary, compared with passive training programs, error capture, feedback, and active situation-based practice in a simulated environment is a much more effective strategy for raising awareness and increasing the road scanning behavior of older drivers.

The traditional travel demand modeling process is designed to develop relationships among travel characteristics, land use activities, and a simulated transportation system network. The models are calibrated to base year conditions through comparisons of theoretical travel patterns, trip length characteristics, and traffic volumes with similar observed values. For projections of travel to future time periods, the assumption is made that the established database and relationships, and in particular the traffic analysis zones to which these data are assigned, will be the same for the forecast period as at the time of calibration. The primary objective of this research was to establish criteria for decision-making regarding which zones should be subdivided and to develop a technique for deciding how to subdivide. The eventual goal of the research was to improve travel demand forecasts. This research investigated whether travel demand estimates can be improved through the reconfiguration of traffic analysis zones for the future year projection period and through an associated restructuring of the centroid connectors for the future network. Similar subarea techniques have been applied by others in attempts to develop more accurate traffic data for project-specific needs. The subarea approach suffices for project needs, but the adjustments and modifications are not usually fed back into the modeling process. Therefore, the time and effort expended in obtaining project data are not applied to the improvement of the overall travel demand forecasts on a study-wide or regional basis. Under these conditions, system planning efforts and corridor analyses do not receive the benefit of updated information or system refinements. The findings of this research indicate that there is little improvement in the travel forecasts as a result of the subdivision of 9 zones in the New Castle County model into a total of 23 new units. Link-by-link comparisons of traffic assignments based on the New Castle County original 228 zone system and the modified 242 zone system reveal no significant improvements. One reason that can be identified is the detail of the simulated roadway system, which nearly duplicates the existing street and highway network and leaves little opportunity to provide alternative travel routings resulting from the creation of small traffic analysis zones.

Thesis (Ph. D.)--Rutgers University, 2007. / "Graduate Program in Civil and Environmental Engineering." Includes bibliographical references.

Thesis (Ph.D)--Aerospace Engineering, Georgia Institute of Technology, 2009. / Committee Chair: Mavris, Dimitri N.; Committee Member: Baik, Hojong; Committee Member: DeLaurentis, Daniel; Committee Member: Lewe, Jung-Ho; Committee Member: Schrage, Daniel. Part of the SMARTech Electronic Thesis and Dissertation Collection.

Thesis (M.S.)--Northwest Missouri State University, 2008. / The full text of the thesis is included in the pdf file. Title from title screen of full text.pdf file (viewed on July 1, 2009) Includes bibliographical references.

Thesis (Ph.D)--Civil and Environmental Engineering, Georgia Institute of Technology, 2008. / Committee Chair: Adjo Amekudzi; Committee Member: Catherine Ross; Committee Member: Josias Zietsman; Committee Member: Michael Meyer; Committee Member: Randall Guensler.

<p> Convective weather and other constraints create uncertainty in air transportation, leading to costly delays. A Ground Delay Program (GDP) is a strategy to mitigate these effects. Systematic decision support can increase GDP efficacy, reduce delays, and minimize direct operating costs. In this study, a decision analysis (DA) model is constructed by combining a decision tree and Bayesian belief network. Through a study of three New York region airports, the DA model demonstrates that larger GDP scopes that include more flights in the program, along with longer lead times that provide stakeholders greater notice of a pending program, trigger the fewest average arrival delays. These findings are demonstrated to result in a savings of up to $1,850 per flight. Furthermore, when convective weather is predicted, forecast weather confidences remain the same level or greater at least 70% of the time, supporting more strategic decision making. The DA model thus enables quantification of uncertainties and insights on causal relationships, providing support for future GDP decisions. </p>

Thesis (Ph.D.)--University of Texas at Arlington, 2009.

No description available.