Journal of Software:2015.26(2):380-389

(西北工业大学 计算机学院, 陕西 西安 710072 ;国家数字交换系统工程技术研究中心, 河南 郑州 450000;北京控制工程研究所, 北京 100190)
Statistical Model Checking of Cyber-Physical Systems Control Software
SHAN Li-Jun,ZHOU Xing-She,WANG Yu-Ying,ZHAO Lei,WAN Li-Jing,QIAO Lei,CEHN Jian-Xin
(School of Computer Science and Technology, Northwestern Polytechnical University, Xi'an 710072, China ;National Digital Switching System Engineering and Technological Research Center, Zhengzhou 450000, China;Beijing Institute of Control Engineering, Beijing 100190, China)
Chart / table
Similar Articles
Article :Browse 3810   Download 4678
Received:July 15, 2014    Revised:October 31, 2014
> 中文摘要: 信息物理融合系统常采用嵌入式实时多任务系统作为其控制软件,这类软件的并发和非确定性给验证带来了困难.提出了一种利用统计模型检验技术分析多任务系统的功能正确性的方法.该方法构造的时间自动机模型以模块化的方式描述了实时多任务系统中的主要成分,包括实时操作系统、周期性任务、偶发任务、共享资源以及物理环境,能够展现多任务系统的细粒度的运行过程及其对物理环境的实时响应.应用该方法分析了玉兔号月球车控制软件的一个早期版本,发现了系统运行中出现的一个特殊错误,识别了实际系统出现错误的条件,再现了出现错误的场景.
Abstract:Cyber physical systems (CPS) typically employ real-time multitasking systems as their control software. This paper proposes an approach to formally analyzing such control software using statistical model checking of UPPAAL. The main contribution of this study is a model in timed automata which modularly describes the major components of a multitasking system. The model supports the analysis of timing-related functional properties as well as schedulability analysis, and can easily be adapted and extended for verifying different properties of various multitasking systems. A case study on an early version of the Chinese Lunar Rover control software shows that the proposed method is able to track down undesired behavior in real-world industrial CPS.
文章编号:     中图分类号:    文献标志码:
基金项目:国家自然科学基金(61472327) 国家自然科学基金(61472327)
Foundation items:
Reference text:


SHAN Li-Jun,ZHOU Xing-She,WANG Yu-Ying,ZHAO Lei,WAN Li-Jing,QIAO Lei,CEHN Jian-Xin.Statistical Model Checking of Cyber-Physical Systems Control Software.Journal of Software,2015,26(2):380-389