Projekte

HFT auf Facebook      HFT auf Twitter   HFT auf Youtube   HFT und Snapchat

Robust algorithms for the calculation of distances between large, moving triangle meshes <<

Akronym RASAND
Funding Bundesministerium für Bildung und Forschung (BMBF)
Request for Proposal FHprofUnt Forschung an Fachhochschulen mit Unternehmen
Supporting Organisation Arbeitsgemeinschaft industrieller Forschungsvereinigungen Otto von Guericke e.V. (AiF)
HFT project director Prof. Dr. Nicola Wolpert
Project Staff Dipl.-Math. Rainer Erbes, Dipl.-Math. (FH) Anja Mantel
Project partners Teraport GmbH, Aschauer Straße 32, 81549 München und Institut für Informatik,Johannes Gutenberg Universität Mainz, Staudingerweg 9, 55099 Mainz
Duration 01.08.2010 - 31.07.2013
Project description This project is motivated by challenges faced in the field of vehicle design. In particular, the spatial validation of component geometries, motion patterns (e.g. crash behaviour), fitting paths and serviceability needs to occur at an early stage in the design process. The aim of this project is therefore to reliably calculate safe distances between moving parts. The parts under investigation move through the space around them over time and are modelled as triangle meshes which usually consist of many hundreds of thousands of triangles. The fact that part orientation and positioning is known at discrete points in time makes it possible to describe the movement of the part. Up to a hundred thousand configurations may be possible. With this in mind, participating researchers aim to identify the triangles in the two objects and the moments in time which describe the point at which the spacing between the two parts is minimal and therefore likely to remain within the specified tolerance. The special challenge is to develop algorithms able to cope with the extremely large volumes of data involved. The exploitation of the parallelism offered by the increasingly widespread use of multi-core processors and the integration of stream processors into graphics cards is indispensable if this goal is to be achieved. In addition to run-time behaviour and memory use, the reliability of the techniques developed will also play a decisive role in whether or not they can be transferred to industrial scenarios. This project therefore aims to develop software that efficiently and reliably calculates whether or not component spacing complies with the specified tolerance range (i.e. compliance must not go undetected, for example as a result of errors in floating-point arithmetic).