Research project CP/34 (Research action CP)
Context
During the last decades mobility of goods and persons - notably the share of road transport - has increased significantly. One of the most dramatic and direct effects of the increased mobility on the human environment is the annual occurrence of more than 51,000 accidents with injuries (NIS, 1998). Under these circumstances, each Belgian citizen has a very high risk of being confronted with several accidents, of which at least one is with physical injury.
The competences on the matter are also very dispersed: vehicle safety is a European competence closely related to technological changes in the industry; fiscal measures and traffic regulations are a federal competence, public works is divided between the regional-, the provincial- and the local authorities.
Project description
Objectives
In the previous research programme (Impact of spatial planning on traffic safety, 1998-2000) new methods were developed to locate accidents on the entire network and to define black zones based on statistically sound methods. The explanatory model for traffic safety was based on environment, road and traffic characteristics.
The main objective of this research is to improve that explanatory model for traffic safety, in order to clarify the interactions between safety factors. This requires an examination of utility of new data sources and of innovative analysis techniques. This model will be build on three subresearch projects:
- High resolution satellite imagery: development of new causal factors (SADL KULeuven R&D)
- Spatial Data Mining: determination of significant combinations of causal factors (LUC)
- Multi Level Analysis: analysis of interactions at different scales (UCL)
Methodology
One part of the project consists of determining environment and traffic characteristics from high resolution satellite images. The new generation of satellite imagery has spatial resolutions up to 61x61 cm (Quickbird) which makes it possible to discern individual vehicles. By means of specific software and image processing techniques we will try to extract characteristics of vehicles, roads and environment in a semi-automatic way. Subsequently it will be investigated if those indicators take on different values within and outside black zones.
The Spatial Data Mining research part will focus on the accident database. This database contains thousands of localized accidents in Belgium, each accident record having about 65 attributes. Data Mining techniques search for significant combinations of those factors. Once the black zones are located, correlations between the attributes will be explored. Aren’t there any specific correlations for specific accidents or black zones? Is it possible to classify those accidents and zones?
The Multi Level Analysis will explore the effect of different aggregation scales on traffic safety. It can be statistically demonstrated that results obtained at one scale of analysis can not be extrapolated to another scale. Multi level analysis is a statistical method which specifically addresses the problem of the spatial scale of the analysis and the surrounding scales ("nested scales"), by analyzing the accident characteristics on a local scale and at the same time being aware of the differences in the context or environment. Concerning traffic accidents, data are available about individual accidents and individual actors.
Interaction between the different partners
Every research team will elaborate an own analysis technique. Conformity will be brought by two successive iterations. These provisional results of each team will be used as input for following steps in the research of all other teams. During the first iteration the accident database as well as an exploratory statistical analysis of accident causes will be completed. These results are used as input in the second iteration which will refine those results.
Expected results and/or products
One of the results will be a database of localized environmental elements, road infrastructure and accidents between 1991 and 2000. This database will be used by the researchers themselves and via the users (especially the BITS) put at the disposal of road managers and other researchers. The results will be published in specialized journals and presented at international colloquia concerned with road safety.
Partners
Activities
SADL, a division of the KULeuven, is a multidisciplinary research group specialized in processing spatial data. SADL applies specialized technical knowledge in the fields GIS and Remote Sensing on subject matters ranging from hydrology studies to traffic safety.
The Research group "Data Analyse en Modellering" from the University of Limburg (Limburgs Universitaire Centrum) already has experience with Data Mining techniques on economical matters, which they will apply on traffic safety matters. Recently they accommodate the "Steunpunt Verkeersveiligheid" (Centre for Traffic Safety).
The experience of the "Centre d’Analyse Spatiale et Urbaine" of the "Département de Géographie" of the UCL mainly concentrates on spatial statistics and a thorough knowledge of traffic accidents statistics. In the previous research project they developed a new method for defining black zones by means of spatial autocorrelation techniques.
Coordinates/Details
Thérèse Steenberghen (project coordinator)
Katholieke Universiteit Leuven (KULeuven)
R&D - SADL
Vital Decosterstraat 102
B-3000 Leuven
Tel: +32 (0)16 32 13 37; Fax: +32 (0)16 32 97 60
Therese.steenberghen@sadl.kuleuven.ac.be
http://www.sadl.kuleuven.ac.be
Geert Wets
Limburgs Universitair Centrum (LUC)
Departement Bedrijfskunde
Universitaire Campus, Gebouw D
B-3590 Diepenbeek
Tel: +32 (0)11 26 86 49; Fax: +32 (0)11 26 87 00
Geert.wets@luc.ac.be
http://www.luc.ac.be
Isabelle Thomas
Université Catholique de Louvain (UCL)
Département de Géographie
3, place Louis Pasteur
B-1348 Louvain-La-Neuve
Tel: +32 (0)10 47 21 36; Fax: +32 (0)10 47 28 77
Isabelle@ucl.ac.be
http://www.ucl.ac.be
Users Committee
- Didier Antoine - Ministère de l'Equipement et du Transport de la Région Wallonne - Direction du Trafic et de la Sécurité routière - Namur
- Philippe Barette - Ministère de la Région Bruxelles-Capitale - Administration de l'Equipement et des Déplacements - Bruxelles
- Marc Broekaert - SPF Mobilité et Transports - DG Mobilité et sécurité routière - Bruxelles
- Alain Broes - Ministerie van het Brussels Hoofdstedelijk Gewest - Bestuur van de Uitrusting en het Vervoer - Brussel
- Luc Deryck - Ministerie van de Vlaamse Gemeenschap - Mobiliteitscel - Brussel
- Wendy De Weser - Federale Politie - Openbare Orde en verkeer - Brussel
- Erwin Debruyne - Vereniging Vlaamse Steden en Gemeenten - Brussel
- Karin Dubuisson - Ministère de l'Equipement et du Transport de la Région Wallonne - Direction du Trafic et de la Sécurité routière - Namur
- Serge Kinet - Police fédérale - Direction de la banque de données nationale - Bruxelles
- Sylvie Le Maire - Federale Politie - Verkeerspolitie - Brussel
- Jeannine Lees - FOD Mobiliteit en Vervoer - Brussel
- Pascale Lucas - Federale Politie - Brussel
- Gaetan Maarten - Federale Politie - Openbare Orde en Verkeer - Brussel
- Koen Ricour - Federale Politie - Verkeerspolitie - Brussel
- Ward Vanlaar - Belgisch Instituut voor de Verkeersveiligheid vzw (BIVV) - Brussel
Innovative spatial analysis techniques for traffic safety : final report
Steenberghen, Thérèse - Thomas, Isabelle - Wets, Geert ... et al Brussels : Belgian Scientific Policy, 2006 (SP1584)
[To download]
[To order]
Techniques innovatives d'analyse spatiale en matière de sécurité routière : résumé
Bruxelles : Politique scientifique fédérale, 2006 (SP1586)
[To download]
Innovatieve Ruimtelijke Analysetechnieken voor Verkeersveiligheid : samenvatting
Brussel : Federaal Wetenschapsbeleid, 2006 (SP1587)
[To download]
Innovative spatial analysis techniques for traffic safety : summary
Brussels : Belgian Science Policy, 2006 (SP1588)
[To download]
