Report numberRA-2005-59
TitleRisk analysis cycle on Flemisch motorways
SubtitlePart II: analysis of the Flemsh road characteristics
AuthorsFrank Van Geirt
Erik Nuyts
Published byPolicy Research Centre for Traffic Safety 2002-2006
Number of pages119
Document languageDutch
Partner(s)LUC (nu UHasselt)
Work packageOther: Infrastructure and space

In this second part of the risk analysis cycle on Flemish motorways, we do a further investigation of the road characteristics. Correlations between characteristics and/or other traffic parameters are not done. The characteristics are studied separately. We start by using ARAN data in order to tell whether specific motorway sections are according to the standards and guidelines, as collected in part 1. Since it is the first time that ARAN data is used for this purpose in Flanders, we also investigate the practicability of it. The bases for this study are pictures and measurements from March 2003.


Missing data is added by using imputation techniques. Some techniques are compared for different road characteristics. For non-discrete data, like lane widths, the linear interpolation technique seems to be the best method. Discrete data, like number of lanes, can be best imputed with the Last Observation Carried Forward method. In this case a known value is kept constant until another value is found.


An important conclusion of this report is the measurement error when using the ARAN data. Several causes exist for this error. For road characteristics like lane width of the lane on which the pictures were taken, the error is very low and acceptable. It is comparable with the errors mentioned by the software supplier. For road characteristics like the shoulder width or the redress lane width, the errors are bigger. After analysis, it seems that the main cause can be found with the measurement procedures. For future analyses, it is thus possible to improve the error and the road characteristic data.


Despite the big errors for some characteristics, we still want to tell something about the road characteristics, based on this database. For that reason we calculate the 95% confidence intervals. So, in the most positive situation for the infrastructure, we can tell that on about 48 km motorway the right-most lane is smaller than 3.50 m.


For the shoulder width, we find a big error. The width of the shoulder changes very much along the motorway, e.g. due to piers of bridges. We get big changes when we calculate the means on road sections of 100 m. The measuring procedure has to be improved definitely. Looking at the most positive situation, we find about 80 km of motorway with 100 m road segments where the mean shoulder width is below 3.00 m, stated in the guidelines.


Based on this database, we can not judge the redress lane. The error is bigger than the measurements. This error can also be reduced a lot by improving the procedure. When taking pictures and measurements on the left-most lane, the error for this characteristic will drop drastically.


Other road characteristics are also calculated and evaluated. For example, distribution of the number of lanes, the density of objects along the roadside, distance between ramps, lengths and widths of ramps.

DownloadPDF icon RA-2005-59.pdf


The Policy Research Centre for Traffic Safety carries out policy relevant scientific research under the authority of the Flemish Government. The Centre is the result of a

cooperation between Hasselt University, KU Leuven and VITO, the Flemish Institute for Technological Research.


Leuven vito