A System-Level Approach to the Design and Development of Space-Based Embedded Systems
- Paper number
IAC-06-D1.P.1.03
- Author
Mr. Andrew Riha, Iowa State University, United States
- Coauthor
Dr. Diane Rover, Iowa State University, United States
- Year
2006
- Abstract
As the Vision for Space Exploration enters the generation of manned missions to the Moon and Mars, an unprecedented need for advanced space-based embedded systems arises. Space-based embedded systems, which manage everything in a spacecraft from attitude control to image processing, require a high degree of fault-tolerance - this arises directly from the extreme conditions that space imposes on the operation of a spacecraft's components. In addition to fault-tolerance, space-based embedded systems must exhibit redundancy in order to guarantee tasks meet their deadlines should a component fail. With routine space missions planned for the near future, real-time space-based embedded systems will need to be rapidly designed with characteristics such as fault-tolerance, redundancy, and reliability in mind. Embedded system development is optimized through a system-level design methodology - high-level models are refined until a complete system specification is obtained. In this manner, advanced embedded systems can be rapidly developed and more time can be devoted to refining and testing the system. In this paper, we examine the properties of fault-tolerant embedded systems that are not currently characterized in system-level models. For example, although system-level design languages such as SpecC support the characteristics of concurrency and hierarchy, redundancy is not currently supported. In addition to defining properties that are not currently characterized, we present modifications to the system-level modeling environment so that new properties are captured and their respective performance metrics are analyzed. Finally, we investigate the effect of our modifications on system-level design models such as the specification model and how those modifications affect the process of model refinement. Through our analysis, we provide a methodology to accelerate and enhance the design and development of space-based embedded systems.
- Abstract document