This manuscript summarizes my research activities since my thesis defended September 2002. Some of my works presented here are completed; others are in progress or still at an exploratory stage. Throughout these years, my works were in the context of software/hardware co-design of intensive signal processing specific Sacs. The complexity of systems targeting such application domain continues to expand, in recent years. Indeed, the growing needs in terms of computing power and memory storage of intensive signal processing applications, makes designing SoCs dedicated to them very tedious, requiring considerable time and effort. Thus, the guideline of my work has always been to provide methods and tools for the design of such Sacs, allowing maximum automation, increasing designer's productivity and reducing time-to-the market of designed systems. So I concentrated my research effort on three main aspects: high level modeling by providing meta-models and profiles based on the MARTE standard; distributed simulation platforms, supporting interoperability between different abstraction levels while allowing accurate energy consumption estimation; and finally production of design tools based on automatic model to model transformations model and MDE approach. Being convinced of the great potential of partially and dynamically reconfigurable FPGAs, I focus more and more my work to target such architectures. Thus, my future work will certainly be in the same direction, particularly in the context of the ANR project FAMOUS. Thus, the main concern of my research will include: modeling (based on MARTE) of dynamic reconfiguration in all its aspects (architecture, application, association, deployment and partitioning), simulation of FPGAs and the estimation of their consumption (to drive architectures exploration), and finally the integration into design tools based on standards (such as MDE and MARTE).
展开▼
