A CO-DESIGN METHODOLOGY FOR DEPENDABLE REAL-TIME SYSTEMS

摘要

A real-time speech'recogniser is designed using the co-design methodology. It is composed of several modules. Each of these modules has its special functions and can be built in software or hardware. There is a large number of possibilities to partition the overall functionality into modules. Thus, a good design methodology must be used to find a good partitioning. In this paper, we use the co-design methodology to partition the hardware and software of a real-time speech recogniser. Three major representative methods are chosen for discussion and performance evaluation. The results show that the co-design methodology can help us to select a good partitioning, whereas we had to give up some designs that could not meet the system specifications. Besides, it can also shorten the design and development cycle and, additionally, development costs can be reduced.