网络嵌入式系统基于事件驱动的代码诊断方法研究

项目名称： 网络嵌入式系统基于事件驱动的代码诊断方法研究

项目编号： No.61472360

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 董玮

作者单位： 浙江大学

项目金额： 82万元

中文摘要： 近年来，嵌入式系统发展迅速，全球嵌入式软件市场每年以超过30%的速度持续增长。特别是以移动智能平台和嵌入式传感网为代表的网络嵌入式系统，具有极为广阔的市场前景。然而，由于体积小、能量受限、交互环境复杂，嵌入式设备经常面临能量、性能等方面的问题。嵌入式代码诊断是解决上述问题的有效方法。本项目将在网络嵌入式系统中展开基于事件驱动的代码诊断方法研究。其具体研究内容包括：1. 基于事件驱动系统事务的抽象、构造与审计方法；2. 基于同构事务时空关联性的异常检测方法；3. 基于属性依赖网络的代码诊断方法。最后，针对智能移动平台和嵌入式传感网进行案例研究，改善现有代码诊断方法的精确性，最终形成统一的基于事件驱动嵌入式系统的代码诊断方法。本项目的研究将对解决嵌入式系统能量、性能等方面的问题具有重要的研究价值和广泛的应用场景。

中文关键词： 网络嵌入式系统；代码诊断；事件驱动

英文摘要： The recent years have witnessed the rapid development of embedded systems. The international market of embedded software is growing more than 30% annually. Networked embedded systems, such as smart mobile platforms and embedded sensor networks, are envisioned to support numerous applications. Embedded devices often experience energy and performance issues due to their small form size, limited energy supply, and the complex interactions with the external environment. Code diagnosis is an effective method to address these issues. This project will conduct research on event-based code diagnosis techniques for networked embedded systems. Specially, we study (1) the abstraction, construction, and accounting of transactions in event-driven embedded systems; (2) anomaly detection method based on spatial-temporal correlation of isomorphic transactions; (3) code diagnosis method based on dependency networks. Finally, we will conduct case studies for smart mobile platforms and embedded sensor networks. The proposed code diagnosis method will greatly improve the accuracy of existing diagnosis methods, provide a basis for a unified code diagnosis framework. The subject of this project has an important research value and a wide range of applications in addressing energy and performance issues for networked embedded systems.

英文关键词： Networked embedded systems;code diagnosis;event-driven