As complexity increases in embedded systems design, there is need for more time for verification purposes. For embedded systems, the only verification that can be done is running test cases, and the number of cases increases exponentially. In order to shorten this verification phase of the design, we propose a methodology to do formal verification of embedded systems. In formal verification no test cases are needed, but an mathematical analysis of the original model and the refined one. At the present, there is no proven formal verification techniques for a system level design of an embedded system. In this article, we present a brief summary of Model Algebra, present two new transformation rules and experimental results in isomorphism checking based on these rules. We test this transformation rules in various scenarios and prove that it is a promising methodology to improve embedded system design.
ResumenCon el incremento en complejidad y miniaturización de los sistemas computacionales actuales, su diseño se ha vuelto cada vez más difícil y lento. Para ello se necesitan modelos óptimos, y desarrollados tempranamente para probar los componentes. Para poder crear modelos automáticamente, se necesitan estructuras de datos y herramientas adecuadas. Si el modelo está bien declarado, con reglas y estructuras definidas, es posible crear un algoritmo que lo genere automáticamente. Para lograr la generación, se definió un estilo de modelaje y una estructura de datos que capturara todas sus características. Luego se implementaron algoritmos en C++ en una herramienta para generar modelos usando como entrada una aplicación en C y especificaciones de la arquitectura del sistema. La salida consistió en un modelo ejecutable del sistema en SystemC. Se seleccionaron dos aplicaciones industriales como prueba: un decodificador MP3 y un decodificador H264. El aporte de la investigación consistió en la definición de modelos adecuados para generar algoritmos que logran producir modelos ejecutables a nivel de transferencia de forma automática y sin necesidad de codificación en otro lenguaje, en una fracción del tiempo requerido. También, se desarrolló una herramienta que genera automáticamente modelos a nivel de transferencia de sistemas computacionales incrustados, útiles para simulación de alto nivel de abstracción para decisiones de arquitectura y desempeño.Palabras clave: Generación automática; modelos a nivel de sistemas; SystemC; sistemas incrustados, ingeniería eléctrica.. AbstractThe design of current embedded computer systems has evolved into a lengthy and expensive process, which needs accurate and early models in order to test its components. In order to generate these models automatically, we need strictly defined data structures and adecuate tools. If the model is well declared, with clearly defined rules and structures, it is possible to create an algorithm that can automatically generate models. The main benefit of automatization is that the designer can move into a higher abstraction plane where he can focus on more important design decisions: architecture, performance, improving his productivity. To achieve automatic generation, we defined a modeling style and data structure that would capture all important characteristics of a system. Then, the generation algorithms were implenented in C++ in a tool that could take a C application, system specifications and would generate an executable SystemC model of the embedded system. We chose two industrial applications in order to test the algorithms: a MP3 decoder and a H264 decoder. The tool generated quickly good quality models and were sucessfully simulated alongside the original C model.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.