En route to safer roads: How road structure and road classification can affect road safety

In the eighties I conducted a literature survey on the interaction between urban planning, road design and road safety.Many years later, I again broached this topic as part of the ʹSafer Transportation Network Planningʹ project (a co-operation between the SWOV and a Canadian Insurance company).In both projects it appeared to be difficult to show a (quantitative) relationship between the planning level and the crashes at an operational level.Also the relationship between road network structure and road safety was difficult quantifying.Furthermore, the Dutch concept of Sustainable Safety requires that the fastest route should coincide with the safest route, another aspect demanding further research.In order to elaborate the issues regarding urban planning, road network structure, route choice, and road safety, I proposed a long-term study ʹRoute choice in road networksʹ.In this project I thought it would be possible to quantify the interactions between these factors by using a simulation model.This would lead me and my fellow colleagues at the SWOV into the world of micro simulation.The study was initially (in the year 2002) going to be carried out by one of my junior colleagues.However, that colleague quite unexpectedly decided to leave SWOV.I had to try and find a replacement, preferably a PhD student.Finding one in the short term was not possible.To compound the problem, my research theme ʹRoad design and road safetyʹ was nearing the end of its four year term and a replacement research subject had not yet started.Combining the two ʹvacanciesʹ (both researcher and subject), resulted in the decision to do the project myself, and thereby taking the first steps towards my PhD project.A review of professors active in the field covered by my project, led me to Martin van Maarseveen as the most promising supervisor.Despite the only relatively large distance between the university in Enschede and the SWOV institute in Leidschendam, a good working relationship was soon established.Although it took some time to get used to each other, we ended up co-operating well.Martin is very diplomatic and provides input in a very subtle way.This requires listening carefully!The study started with an inventory of models that could be useful for our approach.Luc Wismans (consultant at Goudappel Coffeng) was very helpful in providing us with information about this topic.The next step was to choose a micro simulation model.Ronnie Poorterman (consultant at Grontmij) was the first in offering us the S-Paramics model for research purposes.Hans Drolenga (at first as an MSc student, later on as a researcher) managed to make that model appropriate for our study; resulting in our first joint TRB paper.Vincent Kars has gradually improved the application that transforms the output from the model into different types of safety indicators.I would also like to acknowledge and thank the many researchers that have worked on parts of the study: