A 128-node machine will provide the platform for the system. It will be able to acquire images by means of charged-coupled device (CCD) cameras. An existing neural-based measurement system developed at Universita di Roma provides a firm baseline for the work.
The outcome of PATRANS will be a non-intrusive measurement system that can be used for both industrial and research applications, particularly in the aerospace and automotive sectors. Exploiting a parallel platform is expected to improve both functionality and performance in the most cost effective manner.
Project activities include:
optimisation of the neural algorithms for particle recognition and tracking based on existing results from national R&D, together with the adaptation of current Community R&D results where applicable
implementation on a parallel system with optimisation of the hardware image acquisition components
dissemination of the results and establishment of links at a transnational level for possible future cooperation.