一种辐射环境下瞬时故障的软件检测方法  

A Software Approach to Detecting Transient Faults in the Radiation Environment

作　　者：李建立[1] 谭庆平[1] 徐建军[1] 

机构地区：[1]国防科学技术大学计算机学院,湖南长沙410073

出　　处：《计算机工程与科学》2010年第3期115-118,共4页Computer Engineering & Science

摘　　要：空间辐射环境中,大量的宇宙射线经常导致星载计算机出现瞬时故障,这些瞬时故障致使程序执行出现数据错误或者控制流错误。针对瞬时故障导致的程序错误,本文提出了一种软件实现的故障检测算法SITFT,它结合软件复算和标签分析的方法,既可以检测程序运行中的数据错误,又可以有效检测控制流错误。故障注入实验的结果表明,SIT-FT算法在性能开销比源程序增加58%～111%,存储开销增加153%～225%的前提下,使程序执行出现错误结果的情形比源程序减少了49.0%～73.2%。In the cosmic radiation environment , a large number of cosmic rays often lead to the transient faults in onboard computers ,which result in data errors or control flow errors during program execution. Against program execution errors caused by transient faults, a software-implemented fault detection algorithm named SITFT is proposed, which not only detects data errors but also detects control flow errors effectively based on program redundancy and signature analysis. The results of fault injection experiment show chat the misdata in the SITFT program output decrease by 49.0%-73. 2% on the precondition that the execution time increases by 58%- 111 % and the memory cost increases by 153 % -- 225 % than the original program execution.

关 键 词：数据错误 控制流错误 故障检测 故障注入 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]