基于SystemC和POOSL的语音识别系统的系统级建模与性能分析的对比

Contrast of system-level modeling and performance analysis of speech recognition system based on SystemC and POOSL

根据复杂嵌入式系统的系统级建模的需要,提出了一套复杂嵌入式系统的系统级设计流程和建模方法。该方法主要包括了层次化建模方法、Top-Down建模方法、面向对象建模技术和Y-Chart建模方法,并采用提出的系统级建模方法,分别使用SystemC语言和POOSL语言成功地对语音识别系统进行了建模;通过模型的仿真进行了性能分析,并对分别得到的性能分析结果进行了对比,两种语言的系统延迟时间与CPU占用率的性能分析结果基本一致,为建模人员进行复杂嵌入式系统的系统级建模和性能分析提供了指导与参考。

This paper introduced system-level design flow and modeling methods for complex embedded systems. The method included the hierarchical modeling methods, the Top-Down modeling methods, object-oriented modeling technology, and the Y-Chart modeling methods. By using the system-level modeling method, the speech recognition systems which based on Sys- temC and POOSL respectively have been modeled successfully, and analyzed the performance by using the simulation methods. By contrasting the performance results which based on the different modeling language, the same performance results include the system delay time and CPU consumption rate have been gotten, and the paper provides an effective reference in modeling and performance analysis with complex embedded systems for designers.