Haalbaarheidsstudie naar een database voor onderzoek op het gebied van wegen en verkeer.

SWOV’s new research programme started in 2007. This research programme consists of ten sub-programmes, and Roads and Traffic is one of them. Research database is a project within this sub-programme. In this project, data needs to be acquired for study of the effects of road and traffic characteristics on road safety, which is the subject of the project Advancing Sustainable Safety – Layout (ASS – Layout). Knowledge of these effects is relevant for researchers, road authorities and policy makers. Although many road authorities maintain databases with road and traffic characteristics, there is no suitable research database in the Netherlands at this moment. Therefore, there is a need for a database with road and traffic data, which is specifically meant for road safety analyses and policy, and related applications. Internationally such databases already exist. The United States developed the Highway Safety Information System (HSIS) in the 1980s. In this system, many crashes, road and traffic data are recorded for nine states and two urban centres. Every year, the data is updated. Another kind of database is MOLASSES in the United Kingdom. This database contains information on measures taken by local road authorities for local roads which are meant to address a road safety problem. MOLASSES has not been maintained for a number of years and is therefore not up-to-date anymore. The HSIS meets SWOV's demands of a research database quite well. The structure of the Dutch database Wegkenmerken+, with road characteristics and traffic volumes, strongly resembles the HSIS, but contains far fewer road characteristics and is not complete by far. That is the reason why Wegkenmerken+ has not yet been found suitable as basis for the desired research database. In the United States, a list has been drawn up of 180 road and traffic characteristics that are considered the minimum required for road safety research: the Model Minimum Inventory of Roadway Elements. SWOV aims at a database with a more limited number of road and traffic characteristics. The database should contain data on the quantity and composition of traffic. The desired road characteristics should describe the function (for example type of road and speed limit), the alignment (for example visibility, bendiness and hilliness) and the cross section (for example number of lanes, road width, shoulder layout). The database must also contain data on intersections, such as the type (roundabout, priority intersection et cetera), the number of legs and the presence of bicycle and/or pedestrian facilities. Unambiguous definitions and encodings should be drawn up for all characteristics. In the project ASS – Layout, SWOV intends to develop accident prediction models. In the project the choice was made not to include all sorts of road characteristics in the models as explanatory variables, but to set up models. for homogeneous groups of roads. Such a group should contain sufficient road sections to make reliable analyses. Therefore it is advisable that the database contains all roads in the Netherlands. This is very innovative, and therefore, initially, only accident prediction models are developed for rural distributor roads. These roads have already been inventoried for the provinces of Gelderland and Drenthe. This inventory can be considered a pilot study for the feasibility of the desired research database. In this pilot study, the road characteristics were collected manually, which is very labour-intensive. Therefore, direct data input into a GIS is recommended for following occasions. A possibility for collecting data is to request road authorities to hand over their databases and to merge them into one consistent unit. Furthermore, the usefulness of new developments, such as Google Earth, panoramic pictures of CycloMedia and micro-simulation models can also be taken into account. After data has been collected, it is important to perform quality checks. At this stage it should be checked if certain data are not inconsistent or that there are strange differences between road or traffic characteristics of adjacent road sections. When the database is updated on a yearly basis, it can be checked if there are unusual differences between two successive years