Sobhani, Amir Developing a modelling framework to assess safety performance of intersections Road crashes have become one of the most important health issues in today’s society. The number of people killed in traffic crashes in the world each year has exceeded one million people. It is expected that increasing numbers of motor vehicles in many countries will result in a sharp increase in the number of fatalities in traffic crashes to the extent that they may exceed two million injuries by the year 2020. In order to ensure that this does not occur, efforts should be made to manage mobility and safety. Safety evaluation is essential part of safety management. Studies of safety evaluation show that researchers have approached the understanding, analysis and prediction of safety events in many ways. One of these is through the use of mathematical models of safety situations. A literature review of safety models revealed that safety modelling has been approached from three different perspectives. These are the transportation, crash analysis and medical perspectives. Transportation researchers have utilised statistical analysis to investigate the relationship between crash outcome and road, traffic, and environmental factors. Another group of transportation researchers used micro-simulation models to estimate the number and severity of conflicts based on road, traffic and environmental factors. Researchers from crash analysis perspective used statistical analysis and Newtonian Mechanics to explore the relationship between crash severity and crash characteristics such as the mass of bullet vehicle and target vehicle, ∆V of the crash, energy absorption of the colliding vehicles, impact speed, impact characteristics, angle of the crash and equivalent barrier speed. Medical scientists focus on the human body, investigating the level of severity on different regions of the body. Detailed human body characteristics were taken into consideration to measure the level of severity for each region of the body. Statistical analysis is used to study a relationship between occupant injury severity and human body characteristics. In summary, researchers have used micro-simulation models, statistical analysis, and numerical models using Newtonian Mechanics to estimate number of conflicts, severity of conflicts, number of crashes, and severity of crashes. The first three output variables (i.e. number of conflicts, severity of conflicts and number of crashes) have only been considered by transportation researchers. Crash severity is the common output variable used in all three research perspectives to assess road safety performance. All these output variables play a crucial role in road safety assessment. However, the research carried out previously estimates only one or two of these output variables and omits the others in road safety assessment. A comprehensive safety assessment model should estimate all these variables in one framework in order to improve the understanding of variables influencing traffic safety. Moreover, it improves safety modelling by combining these three safety modelling perspectives. The broad aim of this research is to develop a theoretical framework for intersection safety evaluation. In order to develop the theoretical framework first we define Safety Analysis Chain (SACH). The SACH diagram shows the relationship of a chain of events taking place and making a road location unsafe. The five events in the SACH include “traffic flow”, “conflicts”, “severe conflicts”, “crashes” and “severe crashes”. Then a theoretical method is outlined to quantify the five events of the SACH. The developed theoretical method consists of three components outlined below: • The first component of the framework is to estimate the number and severity of conflicts using a micro-simulation model. This component estimates the first three events of the SACH. Inputs into the first component are the geometry and traffic characteristics of the road system. The output of the first component of the simulation framework is the characteristics of the serious conflicts which have been generated in the micro simulation model. • In the next step the link between “severe conflicts” and “severe crashes” is modelled. In order to model this link, the risk of being involved in a severe crash is estimated using conflict characteristics. The second component of this framework is the measurement of the potential injury severity of each simulated conflict. The characteristics of the simulated conflict are used as input. These are the output of the first component of the simulation framework. A three-step modelling approach is used to estimate the potential injury severity of conflicts. In the first step, expected speed change of the subject vehicle (∆Vs) in the crash is estimated given the condition that the conflict leads to a crash. The expected ΔVs is estimated using conflict characteristics and driver’s reaction before crash. The second step is to utilise Newtonian Mechanics to estimate the magnitude of kinetic energy applied to the subject vehicle. In the third step of this component, the expected Injury Severity Score (ISS) of the conflict is measured using estimated kinetic energy of the subject vehicle and the impact type of the expected crash. • The third component of the framework is to estimate average number of crashes and consider this value as the output of “crashes” event in the SACH. The average number of crashes is calculated using either the historical crash data of the road location or the available crash prediction models. The final output of the framework is the safety level of the studied road location. The safety level of the road location is measured using a road safety index (RSI) derived based on the output of the three components of the framework. The RSI is derived based on the definition of crash severity risk (CSR) which is the product of average expected severity and the probability of crash. The value of average expected severity and the probability of crash can be calculated using the quantified events of the SACH. The developed road safety index was applied to assess the risk of turning behaviour at a signalised intersection in Melbourne, Australia. The results of the model were compared to the historical crash data of the intersection. The results of the case study showed the measured risk of being involved in a severe crash for investigated manoeuvres was consistent with the historical crash data of the intersection. This study contributes to knowledge in a number of directions, these are outlined below: • This study contributed on road safety modelling through linking the main safety modelling perspectives to improve road safety understandings and modelling. • An other main contribution of this study is to integrate the preceding safety research perspectives through developing a theoretical framework to model the link between all the SACH events. • In this study, a combination of Newtonian Mechanics and statistical models is incorporated into the micro simulation model to estimate the potential injury severity of simulated conflicts. The developed framework improves the simulation based safety performance assessment by considering the risk of crash severity for different conflicts. • This is the first study investigating the relationship of ΔV, kinetic energy and the injury severity score of the crash to assess roads safety level. The developed modelling framework uses a combination of micro-simulation, statistical and numerical analysis to integrate the conflicts with crash severity to provide a better assessment of road network safety. Given that, at this point, there are no microscopic traffic algorithms reliable enough to replicate accident occurrences, these models applied after the conflict has been detected from micro-simulation represents an approach to improve the usefulness of microscopic simulation for safety assessments. Crash kinetic energy;ethesis-20130910-143240;Surrogate safety measures;Simulation of safety;Occupant injury severity;thesis(doctorate);Crash severity;Micro-simulation;Traffic safety;Road safety index;1959.1/898149;Intersection safety evaluation;monash:120120;Restricted access;2013;Statistical modelling 2017-02-15
    https://bridges.monash.edu/articles/thesis/Developing_a_modelling_framework_to_assess_safety_performance_of_intersections/4657420
10.4225/03/58a4e592827e5