一种以人为中心的风险评估算法

从人为差错的辨识、概率计算和后果量化三个方面讨论人因事件风险评估的流程,并针对这三个关键问题设计相关的解决方案。针对人为差错辨识问题,设计一种统一的人为差错基本分类框架,作为差错辨识过程的模板库;针对人为差错概率计算问题,提出首先计算人为差错总体概率,然后结合历史事故统计资料计算具体差错模式发生概率的新方法;针对后果量化问题,按照先定性后定量的原则,设计一种后果量化值的确定方法。     

行为模式支持下的人为差错分析方法研究

人为差错严重影响系统安全。为全面辨识分析人为差错并找出合适的规避措施,提出一种行为模式支持下的人为差错分析方法。该方法提出“认知过程是场景相关的”观点,建立不同任务场景下可能对应的认知过程。为此,引入行为模式作为两者之间的桥梁。设计一个行为模式分类器,可以确定不同任务场景下可能对应的行为模式;分析不同的行为模式可能对应的认知过程;在此基础上分析不同行为模式下可能对应的差错模式及其成因,并讨论相应的差错规避措施;通过实例演示以上分析过程。     

空中交通管制中的人为差错探讨

近年来,航空事故发生频繁,纵观历年的事故,人为因素导致的航空事故呈现一个上升而主要的态势,做好对于空中交通管制的人为差错问题研究,降低人员因素带来的安全事故成为本文探讨的重点,以更好的方式来提高人员综合性能力,确保空中交通管制的安全。     

DES算法DFA攻击的差分差错位置判别方法

利用DFA攻击方法的1比特差分差错模型,分析了DES算法的1比特差分差错所处的位置与密文差分之间的关系,并对分析结果进行了实验测试。利用分析结果,可以根据密文差分的特点有效地判断1比特差错所处的位置,进而顺利地进行DFA攻击。     

基于灰狼算法的民航维修人为差错评价模型

随着民航运输业的迅猛发展,航空运输量和排班量大幅度增加.航空器在可靠性和安全性等诸多方面都有了大幅度提升.由机械故障导致的安全事故比例从80％下降到了20％,而维修过程中的人为差错占比却直线上升,成为影响民航安全、飞行安全及运行成本的重要因素.因此,民航业对于人为差错备受关注.为了降低民航维修中人为差错的发生几率,提高...     

基于本体的空管人为差错分析研究与应用

为实现空管人为差错致因及危险等级语义分析,开发人为差错本体分析软件。通过空管运行人为差错(HeraJanus)及人为差错预测(Hera-Predict)手册中获取的领域知识构建领域本体;结合事故调查报告知识创建存储本体,建立Hera-Janus本体知识库,根据类与个体之间的关系定义Jena推理规则。通过Eclipse平台中的Jena应用调用Pellet推理机分析实际案例,在人机交互界面给出差错类型及危险程度,其结果表明了空管事故智能化分析的有效性和可行性。     

基于块划分的空时域自适应差错掩盖算法

视频通信是多媒体通信的主要应用方向之一,它在人们信息交流中起着重要的作用,但是视频信息携带的信息量非常大,在传输之前必须要对视频信号进行压缩编码以降低码率,压缩后的码流对传输差错十分敏感,使得视频信息传输过程中面临不可避免的差错和损失。差错掩盖技术作为视频通信系统中抵抗差错的最后一道防线,起着不可或缺的重要作用。针对现有差错掩盖算法的不足,提出了将结构相似性(structural similarity index,SSIM)失真评价准则和边界匹配准则相结合寻找最佳替代块的匹配准则,并在此基础上,提出了一种基于宏块划分的空时域自适应差错掩盖算法。实验表明,利用该算法进行差错掩盖后的视频序列不仅具有更高的客观质量,而且掩盖后的视频视觉质量也更符合人眼的主观感受。对各算法的平均解码帧率的调查结果显示,在权衡掩盖效果和时间复杂度后,该算法优于其他对比算法。     

分布式控制系统并行通信的可靠性设计与差错检测纠正

本文研究了分布式控制系统并行通信的可靠性设计与差错检测、纠正技术，分析了干扰信号的来源与抑制方法，叙述了传输协议的差错检测与纠正原理，给出了抗干扰的接口信号传瑞原理困及基本通信流程图。     

基于TRACEr失误辨识方法的电网调度人因风险认知分析

辨识和分析电网调度中的人为失误对于防范和控制人因风险、保障电网安全稳定运行至关重要。人为失误预测和确认方法在复杂系统的设计、评估和运行中已得到了广泛应用。结合电网调度业务知识和认知心理学模型,对认知差错追溯和预测技术(Technique for Ret-rospective and Predictive Analysis of Cognitive Errors,TRACEr)进行了分析和改进,研究了其在电网调度人为风险分析中的应用。运用提出的人为失误分析方法,对电网调度中的人因风险案例进行了分析,结果表明,基于TRACEr失误辨识方法能较为全面地分析电网调度员的人为失误,并为失误的补救和防范提供有效的改进措施。     

CRC差错校验法在IBM/PC与8031单片机无线通讯中的应用

通过对不同校验方法的分析，提出了适合ＩＢＭ／ＰＣ与８０３１单片机无线通讯差错校验的ＣＲＣ方法并给出了其实现的软件流程图     

打磨机器人误差建模与参数辨识

简要介绍了一种实用的机器人位姿误差建模方法一摄动法,并利用此方法建立了打磨机器人位姿误差模型,对利用三坐标测量机测量的误差数据进行了建模补偿。误差参数辨识中,在传统的九线法的基础上,提出了单点法测量机器人的误差参数,克服了九线法中三测量点位置误差对测量结果的影响,提高了测量结果的可信度。实验结果表明,建模补偿后机器人位姿误差平均值降低到初始值的1／5左右,最大值降低到初始值的1／6左右,验证了所建模型的正确性和参数辨识方法的有效性。     

机器人机构的误差位形模量的概念及在精度分析和综合中的应用

本文在机器人误差分析影响系数法的基础上引入机器人误差位形模量的新概念.文中给出了误差位形模量的定义,讨论了其特点和计算公式.并以3R 机器人机构为例讨论了其在精度分析和综合中的应用,提高了机器人的使用精度与设计精度.     

Post-error expression of speed and force while performing a simple,monotonous task with a haptic pen

Control errors often occur in repetitive and monotonous tasks, such as manual assembly tasks. Much research has been done in the area of human error identification; however, most existing systems focus solely on the prediction of errors, not on increasing worker accuracy. The current study examines force responses before, during and after errors in a simulated assembly line task in order to determine an optimal feedback system for error reduction. Confirming previous findings, enhanced movement speed and reduced force occurred before erroneous trials and slowing occurred after erroneous trials. Given the results, we suggest a haptic feedback system which stimulates users to exert increased force levels after completing an erroneous task in order to increase degree of control and re-build worker confidence and thereby reduce overall error rate.     

Measuring, modeling, and augmenting reliability of man-machine systems

This paper discusses analytical and empirical trends in dealing with human reliability. It reflects on the nature of human error and man-machine reliability, and poses some serious problems of definition which affect numerical results. Techniques for enhancing man-machine system reliability are presented.     

l_1系统辨识中的代数算法及其Worst－case误差

讨论单输入单输出、离散时不变因果系统的ｌ１系统辨识问题。首先提出基于代数方法的代数算法，并分析了该算法的特点；然后估计其Ｗｏｒｓｔ－ｃａｓｅ误差，并证明了该算法的收敛性；最后讨论了在某些特殊情况下该算法的相应形式。所给结果是面向鲁棒控制的。     

The validation of three human reliability quantification techniques - THERP, HEART and JHEDI: part 1 - technique descriptions and validation issues

This is the first of a set of three papers reviewing the validity of three Human Reliability Assessment (HRA) techniques used in the UK to predict human performance in high risk industries. The techniques are used to determine the risks inherent in such industries due to human error, and also the benefit in terms of risk reduction by having human operators in the system. These techniques culminate in a quantitative human error probability for each error or failure identified, predicting how often errors or failed performance will occur. The major question with such an approach is whether such probabilities are accurate. This key validation question is answered in the second paper, which reports the results of a large and independent validation experiment.

Once a technique achieves some degree of predictive validity, the next question becomes whether such a technique can consistently produce valid and accurate results. This is called the reliability of the technique, and is dependent upon its consistency of usage by different assessors. Consistency can only be analysed by investigating the detailed usage of the techniques, and such an analysis is reported in the third paper in this series. The advantage of such an analysis is that it can also lead to specification of practical guidance for practitioners, and may lead to derivation of ways to improve the reliability and consistency of usage of specific HRA techniques. This first paper introduces the three techniques themselves, and reviews the major validation criteria and issues which should be considered when trying to determine if such techniques work.     

WANO人误研究平台

针对传统WANO人误研究工作效率低下,流程混乱的特征,研究从WANO人误研究工作实际需求出发,采用面向对象技术设计,实现了一个基于信息系统的WANO人误研究平台.从平台构建的角度阐述了WANO人误研究平台的功能结构、平台框架、WANO人误数据库构建等技术要点,并介绍了平台在WANO人误数据ETL方面的应用效果.     

Factors affecting accuracy in the quality control checking of fresh produce labels: A situational and laboratory‐based exploration

Manufacturing industries often rely on quality control staff to ensure mistakes are detected before products are shipped to customers. Undetected errors can result in large financial and environmental costs to packaging companies and supermarkets but the contributors to such error are underexplored. The research reported in this paper investigated human error in the quality control checking of information displayed on the labels which accompany packaged fresh produce. Initial work sought to understand the demands of label‐checking in the packhouse environment, through interviews with key quality control staff, in situ observations, and the study of historical error data held by a fresh produce packaging company. This study highlighted the dynamic and cognitively challenging environment in which label‐checking occurred, while the historical error data indicated both the scale of the packhouse's work and the infrequency of error occurring. In a separate strand of laboratory‐based research, experienced and novice label‐checkers were presented with a simulated label‐checking task and a battery of computerized and pen‐and‐paper tests. These tasks were administered to determine whether cognitive abilities could predict label‐checking accuracy in a controlled laboratory environment. Stronger abilities in two cognitive processes (information processing speed and inhibition) predicted greater overall accuracy and higher detection of labeling errors. In identifying potential contributors to human error in the quality control checking of product labels both in situ and in the laboratory, the results are relevant to manufacturing, wherever information is printed on labels, especially when labeling processes depend upon human data entry and human quality control checking.