The PRIME project stands for Predictive Rendering in Manufacture and Engineering and aims to establish an innovative training network where PhD students develop skills and protocols required for the industrial application of predictive rendering technologies. An ITN network in this field contributes significantly to maintaining and increasing the competitive advantage of the European region in this respect by training young researchers in a promising, future-oriented and research-driven application area. This will allow the establishment of long-term collaborations and permanent, structured training programs between partner organisations to conduct research on several important engineering problems in this field. The application areas of these rendering systems include product design, architecture, sensor system calibration, autonomous vehicle systems and manufacturing control.
In the DFKI Department Agents and Simulated Reality (ASR), work is being done on the one hand to transfer advances from credible rendering applications to general electromagnetic waves. This requires algorithmic improvements to handle effects such as ray diffraction, and the resulting simulations must be compared with ground-truth measurements. On the other hand, the degree of physical accuracy required to successfully train machine-learning approaches from synthetic data is being investigated. A key challenge is to create sufficiently large data sets, requiring high-performance implementations, for example by implementing the required algorithms and methods on a GPU. In addition, algorithms need to be selected or developed to capture ripple effects, complex indirect lighting or other challenging effects.
Partners
Univerzita Karlova (CZ), Luxion APS (DEN), University College London (UK), Danmarks Tekniske Universitet (DEN), Linkopings Universitet (SWE), Universidad de Zaragoza (SP), Imperial College of Science Technology and Medicine (UK)
