A model-based design space exploration for embedded image processing in industrial applications

The design and evolution of today´s embedded systems face an increasing complexity both in the applications and in the platforms that support them. The use of complex platforms means that the engineers need to make non-trivial decisions during the design phase including optimized allocation, binding and scheduling. This situation creates a need for better ways to manage project complexity, and Model Driven Engineering is a possible solution. In this work, we propose an extension of a Model Driven Engineering methodology (i.e. HIPAO) to help manage complexity, which includes an automated Design Space Exploration phase that uses Genetic Algorithms to find the best hardware/software partitioning for a given application. We apply the HIPAO methodology to a case study of an autonomous industrial inspection system.