论文部分内容阅读
对复杂嵌入式系统的端到端信息流进行延迟分析是一种有效的实时性评估方法.体系结构分析与设计语言(AADL)是描述复杂嵌入式系统的标准语言,其中端到端流描述组件间的通讯.目前针对AADL模型中端到端流的延迟分析,手工方法能够深入剖析流语义,分析精确度高,但耗时且低效;自动化方法虽有较高的效率,但在延迟属性的覆盖度及语义精细度上都远远不足,导致分析结果精确度低.另外,这两类方法多只关注于最坏情况等典型场景,而无法分析不确定因素的影响.本文提出一种基于时间自动机的端到端流延迟分析方法,首先总结端到端流延迟的影响属性,并建立延迟属性的元模型,在此基础上提出面向流延迟分析的时间自动机模型生成方法,通过对时间自动机的仿真实现流延迟的分析.最后通过案例说明了该方法能够正确表达流的传输语义及延迟属性语义,即有足够的表达能力;展示了方法能灵活分析多样交互场景以及随机时间因素,即有灵活的分析能力;另外仿真过程的状态变迁及时间变量变化过程也为改进设计模型提供依据与建议.
Delay analysis of end-to-end information flow in complex embedded systems is an effective real-time evaluation method.Analytic Architecture and Design Language (AADL) is a standard language for describing complex embedded systems, in which the end-to-end flow description component Between the current AADL model for the end-to-end flow delay analysis, manual methods can analyze the flow semantics in-depth analysis of high accuracy, but time-consuming and inefficient; Although the automation method has a higher efficiency, but in the delay attribute Coverage and semantic precision are far from enough, resulting in low accuracy of the analysis results.In addition, these two methods mostly focus on the typical scenarios such as the worst case, but can not analyze the impact of uncertainties.This paper presents a Based on the analysis of end-to-end flow delay in timed automata, firstly, the influence properties of end-to-end flow delay are summarized and a meta-model of delay attribute is established. Based on this, a time-automata model generation method based on flow delay analysis is proposed, Finally, a case is given to illustrate that this method can correctly express the transmission semantics and delay semantics of the stream, that is, there is enough Communication skills; demonstrated flexibility analysis methods and diverse interaction scenarios and random time factor, that is flexible analysis capabilities; additional state simulation process of transition and change process also provides a variable time basis and suggestions for improving the design model.