Embedded system must perform progressively more complex functions, which requires a high level of embedded intelligence. This rise in intelligence forces a rise in the level of information processing that the embedded systems carry out. This, in turn, gives rise to a significant increase in complexity of the hardware and the software. Due to this, the conception and design of embedded systems for control and/or digital signal processing applications in real-time has to face two major points, in many cases opposed to each other. On one side, a higher level of intelligence is demanded day by day to the systems, and, on the other side, these systems have requirements of real-time operation and reliability difficult to meet in complex systems.
The way to be explored is a) to exploit high-level cognitive architectures in embedded systems and b) to develop scalable architectural models that allow maintaining properties of intelligence and reliability when scaling down in a System-on-Chip.