Fault Detection and Board-Level Diagnosis by Signature Analysis

通过特征分析进行故障检测和板级诊断

• 批准号: 8813748
• 负责人: Mark Karpovsky
• 金额: --
• 依托单位: Trustees of Boston University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1989
• 资助国家: 美国
• 起止时间: 1989-01-01 至 1991-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8813748&HistoricalAwards=false
• 关键词: Fault; Detection; Board; Level; Diagnosis

This is a collaborative effort between Professors M. Karpovsky at BostonUniversity and D. Pradhan at the University of Massachusetts. The proposedinvestigations are two.fold. The first concerns fault detection by acompression of test responses. Errors (manifestations of faults) are definedas the component.wise exclusive OR between actual test responses and fault.free responses. Faults are masked if the observed signature (compressed test response) is equal to the reference. Thus, depending on the statistics of theerrors, an optimal compressor is the compressor that minimizes a probabilitythat the test responses are compressed into the reference. The design ofoptimal compressors when the distribution of ones contained in error sequences is known for a given circuit under test is being investigated. This work isalso based on single error correcting codes over a finite field. Second isboard . level diagnosis by signature analysis. The objective is to analyzetradeoffs between the hardware complexity of the redundant chips on the boardand test time. Two techniques, based on compression of test responses, arebeing pursued. These are; space . time compression of test responses, anextension of previous research in space compressors; and board diagnosis byHamming codes over a finite field. Professors Karpovsky and Pradhan, are outstanding researchers with solidrecords of accomplishment over the past decade, and that they are very likelyto continue first class research. The research topic, test response analysis, is important to industry, and is a high.priority category within the Design,Tools and Test Program.

这是M教授之间的合作努力。Karpovsky和D.普拉丹在马萨诸塞州大学。建议的调查有两个方面。第一种是通过压缩测试响应来进行故障检测。错误（故障的表现）被定义为实际测试响应和fault.free响应之间的组件异或。如果观察到的签名（压缩的测试响应）等于参考，则故障被屏蔽。因此，根据误差的统计，最佳压缩器是使测试响应被压缩到参考中的概率最小化的压缩器。当已知被测电路的误差序列中的1的分布时，最优压缩器的设计是被研究的。这项工作也是基于有限域上的单纠错码。第二个岛。通过特征分析进行水平诊断。目的是分析板上冗余芯片的硬件复杂度和测试时间之间的权衡。两种技术，基于压缩的测试响应，正在追求。这是：空间。测试响应的时间压缩，对空间压缩机的扩展研究，以及有限域上汉明码的电路板诊断。Karpovsky教授和Pradhan教授是杰出的研究人员，在过去十年中取得了扎实的成就，他们很有可能继续进行一流的研究。测试响应分析这一研究课题对工业界很重要，是设计、工具和测试程序中的一个高优先级类别。