EVU XXIII 2014-22 Automated crash computation of passenger car accidents based on the GIDAS database


Michael Wagner, Lars Hannawald, Henrik Liers
Institute for Traffic Accident Research – University of Technology Dresden

Abstract

The accident reconstruction programs currently available on the market allow the user to manually reconstruct a real traffic accident using simulation. Most often pictures of the accident scene and, If available, eye witness statements are used within the reconstruction process. Considering the time and resources needed, this kind of reconstruction takes a lot of effort.

For the estimation of the benefit and effect of innovative DAS (Driver Assistance Systems) on the collision positions and by association on the accident severity, together with the economic benefit, it becomes necessary to simulate and evaluate a variety of virtual accidents with different start values (e.g. initial speed, yaw rate). Only an automated crash computation can be considered for this task.

Within several projects the VUFO GmbH has developed a new method for an automated crash computation based on data from the GIDAS (German In Depth Accident Study) database.

The automated crash compuation is based on standardized voxel-based models of vehicles with specific EES-values for each voxel. These virtual vehicle models are grouped into several categories based on the vehicle deformations obtained from the GIDAS database.

This paper explains the design of the virtual vehicle models, the method of the crash computation as well as exemplary applications of the automated crash computation.

For the first time an automated crash computation of passenger car accidents has been realized.

Using the automated crash computation different tasks within the field of vehicle safety can be elaborated.

This includes, for example, the calculation of specific accident parameters (such as EES or delta-V) for various accident constellations and the estimation of the economic benefit of DAS using IRFs (Injury Risk Functions).

Testo completo della pubblicazione scientifica disponibile ai soli iscritti.


AutoreMichael Wagner, Lars Hannawald, Henrik Liers Institute for Traffic Accident Research - University of Technology Dresden
Pubblicato il
Tipologia