Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/90266
Title: Modeling of pedestrian safety at the macroscopic level
Authors: Su, Junbiao
Degree: Ph.D.
Issue Date: 2021
Abstract: Pedestrians are vulnerable to severe injury and mortality in road crashes. They account for more than 60% of total road fatalities in Hong Kong. It is necessary to promote the understanding of the relationship between pedestrian crashes and possible influencing factors. Therefore, effective countermeasures can be developed to improve pedestrian safety, and more importantly, the well-being of society. The tendency of a person to be involved in a road crash can increase with their amount of travel. This is referred to as exposure. It is necessary to control for the exposure in the estimation of crash risk. A common approach to estimating the vehicular crash exposure is to calculate the traffic flow or vehicle kilometers traveled using traffic count data, which are often readily available. However, accurate and extensive pedestrian counts are seldom available for the estimation of pedestrian crash exposure. In this study, comprehensive population, traffic, and travel data, obtained from multiple data sources, are used to determine the pedestrian crash exposure. In addition, the effects of influencing factors, including travel purposes and transport modes, on pedestrian crash exposure are considered. Then, the association between land use, built environment, population characteristics, traffic conditions, and pedestrian crash risk at the macroscopic level is measured using the Bayesian spatial model. First, the efficient exposure measures for pedestrian crashes are explored, based on the comprehensive travel survey data in Hong Kong. For instance, zonal population, and frequency and time of walking trips are adopted to represent the pedestrian exposure to road crashes. The random-parameter negative binomial regression approach is then applied to measure the relationship between pedestrian crash frequency, exposure, and possible influencing factors. Results indicate that model that applies the frequency of walking trips as the proxy for pedestrian exposure is superior to that using zonal population and walking time. Second, the effects of travel purposes on pedestrian crash risk are examined. Pedestrian crash exposures, represented by the frequency and time of walking trips, are discretized by time of the day and travel purposes. For instance, the trips are stratified into six types, including home, work, school, shopping, dining, and others. Results indicate that the crash risk of "back home" walking trips is amongst the highest. This phenomenon aligns with the distribution of pedestrian crash rate by time of the days, i.e., pedestrian crash rates at noon and in the late afternoon are amongst the highest. This could shed light on the implementation of effective policy strategies that can improve the safety of vulnerable pedestrian groups in specific time periods.
In Hong Kong, 90% of total trips are made by public transport. Walking is the primary mode to get access to public transport services. Safe access to transport facilities is an important issue in sustainable urban development. In this part, pedestrian crash exposures, are categorized by the different transport modes, i.e., metro, bus, light bus, taxi, and private car. Results indicate that pedestrian crash risk is positively correlated to the frequency of walking trips for roadway transport services. However, the association between pedestrian crash risk and frequency of walking trips accessing to metro is not significant. This is indicative to the design and planning of road facilities, i.e., pedestrian crossings and traffic signals, that can enhance the accessibility and safety of public transport. Fourth, a joint probability approach is developed for simultaneous modeling of crash occurrence and pedestrian involvement in crashes, with which the possible correlations among different crash types, i.e., crashes involving and not involving pedestrians, are accounted. Possible influencing factors, including land use, road network, traffic flow, population demographics and socioeconomics, public transport facilities, and trip attraction attributes are considered. Markov chain Monte Carlo and full Bayesian approach then applied to estimate the parameters. Results indicate that crash occurrence is are correlated to traffic flow, the number of non-signalized intersections, and points of interest such as restaurants and hotels. By contrast, population age, ethnicity, education, household size, and road density could affect the propensity of pedestrian involvement in crashes. In conclusion, this study has addressed several fundamental problems, including the estimation of pedestrian crash exposure, effects of travel purposes and transport mode involvement in pedestrian crash exposure, and possible correlations among crashes of different types, for the development of pedestrian crash prediction models. Findings are useful for the design of various transport facilities, road safety education for vulnerable pedestrian groups, and more importantly, accessibility to essential urban services and attractions. With these proposed improvements, a safe and accessible walking environment can be promoted in Hong Kong.
Subjects: Pedestrians -- Safety measures
Pedestrian accidents -- Prevention
Traffic safety
Hong Kong Polytechnic University -- Dissertations
Pages: xv, 144 pages : color illustrations
Appears in Collections:Thesis

Show full item record

Page views

57
Last Week
1
Last month
Citations as of Mar 24, 2024

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.