|Ph.D Student||Sadia Reut|
|Subject||Modeling Drivers' Speed Selection and its Relationship|
to Road Infrastructure Characteristics
|Department||Department of Civil and Environmental Engineering||Supervisors||Professor Shlomo Bekhor|
|Professor Abishai Polus|
|Full Thesis text|
Drivers’ speed selection has been of a great interest to road safety researchers in recent years; average speed and speed distribution were found in various studies to have significant effects on road safety. Various studies were conducted on drivers’ speed selection, focusing on specific aspects, namely infrastructure, environment or driver characteristics. Few studied a combination of different factors that may affect speed selection. Thus, the purpose of this study was to construct a method that allows a combination of factors. For this purpose, a stated-preference survey and a driving simulator (STISIM) experiment were developed and applied in two study phases, exploring the effects of environmental characteristics, driver characteristics, and additional risk/benefit considerations, on drivers’ speed selection.
The first part of the study used a stated-preference web-based survey with a sample of 290 participants. The survey included newly developed scales; estimating driving risks, estimating personal difficulty of performing vehicle-related technical tasks and spatial tasks, and drivers’ own self-assessments. The analysis of the data revealed that latent driver characteristics, such as risk awareness and technical aversion, strongly affect individual drivers’ speed selection. The perceived speed of the average driver also had a significant effect on drivers’ own speed selection.
The second part of the study involved a simulator experiment, in which the roads were portrayed as a four-lane highway with a median barrier, and with low traffic density allowing unconstrained speed selection. Drivers also filled out the survey from the first study. A heterogeneous sample of 111 mostly-experienced drivers drove four trips of about 8 km each, in free flow conditions. Environmental factors included design speed, geometric design of segments, varying traffic speeds, etc. Situations of various risks and benefits have also been presented to drivers, including enforcement, crash risk, and time-saving benefits. Data resolution in the simulator allowed collecting average speed per segment, a total of 9768 road segment observations. 23 crashes occurred in the simulator during the experiment.
Statistical analysis revealed that the overall largest effect on drivers’ speed was when time-saving benefits were presented, followed by increased speeds when the average speed of close-by vehicles was 110 km/hr. Large effects on speed reductions were related to horizontal curves and an increased longitudinal slope, as well as enforcement. Age and gender were also related to drivers’ speed selection. The latent driver characteristics developed by this study were also significant in speed selection.
This study was primarily designed to collect speed selection data in free-flow conditions. The simulator experiment also accounted for any crashes that occurred due to driver behavior. Inspection of the results revealed that a significant amount of crashes had occurred during the experiment, which enabled the estimation of simple crash prediction models. It was found that individual driver speed variance was significant in predicting crash risk for a driver and for a specific trip. Speed deviation from typical behavior of other drivers also construed a higher crash risk. Implications for further speed-selection researches and road safety policies are discussed.