Development and Calibration of Safety Performance Functions for Urban Roads in Flanders, Belgium

Abstract

The primary goal of this doctoral study is to develop safety performance functions (SPFs) for urban roads in Flanders, Belgium. Traffic safety1 in urban areas is a multifaceted issue that requires thorough attention and innovative solutions from planners, engineers, and practitioners. Unlike rural roads, urban streets serve multimodal transport, requiring designers to create a harmonious and comfortable driving environment where longitudinal and transversal vehicular movements can be accommodated and other road users, such as cyclists and pedestrians, can be safely integrated. However, research indicates that urban roads pose elevated crash risks due to high traffic volumes, different types of road users travelling at varying speeds, and the prevalence of daily roadway activities. Recent European Union (EU) data shows that urban areas contribute significantly to overall road crash statistics, resulting in considerable economic and social costs, including injuries and fatalities. Over the past decade (2011-2020), a more aggregated measure has also indicated an increase in the relative proportion of deaths (from 38% to 40%) on urban roads. Meanwhile, the proportion of severe injuries on urban roads has also increased by one per cent (55% in 2011 to 56% in 2020) in the EU. This is also the situation in urban areas of Flanders. These numbers highlight the severity of urban road safety problems, even in the safest regions. By recognising the magnitude of traffic safety problems, governments, transport authorities, and researchers have intensified efforts to enhance safety measures in urban areas. Depending on factors such as available resources, time constraints, and the specific objectives of the assessment, researchers use various methodologies to assess traffic safety and understand crash phenomena. Statistical analysis is the most sought approach to investigate road safety among these methods. Crash prediction models (CPMs) or safety performance functions play a crucial role in several aspects of highway safety assessment, including understanding the causes of crashes, evaluating implemented interventions, and identifying high-risk locations (i.e., hotspots). Moreover, CPMs offer the potential to predict crash frequency and serve as valuable tools for evidence-based policy and urban planning. For this reason, researchers have extensively developed and refined various statistical techniques to analyse crash data and examine the relationship between crash frequency and explanatory variables. However, a literature review about urban SPFs indicated confusing and somewhat contradictory results about the relationship between crash frequency and its predictors on urban road segments and intersections. Despite ongoing efforts to enhance traffic safety, a gap exists in our ability to predict crashes effectively in urban settings. The current reliance on road safety measures is insufficient as urban areas are characterised by complex traffic situations, high vehicle density, and diverse road users. These factors result in challenges that are not fully addressed by the traditional modelling approaches and, therefore, create a need for advanced predictive models to study potential risk factors and contribute to developing targeted safety interventions.