基于支持向量机的高速公路事件检测

针对交通领域中的事件检测(无事件模式和事件模式)模式识别问题,描述了支持向量机(SVM)支持的基本方法,建立了基于线性(linear function)、多项式(polynomial function)和径向基(radial basis function)3种核函数的事件检测SVM模型。采用高速公路路段I-880线圈数据集和事件数据集验证模型,结果发现:无论对于向北、向南或混合方向的事件检测,在3个SVM模型中,SVM(P)检测效果最好,SVM(L)最差。SVM算法具有避免局部最小,实现全局最优化,更好的泛化效果的优点,是高速公路事件检测的一种很有潜力的算法。     

驾驶员自身因素对限速标志识别能力的影响分析

以研究城市道路限速标志下驾驶行为的变化为主要目的,基于认知心理学知识,采用实地观测和问卷调查等手段,研究了驾驶员经过限速标志时的心理及反应过程。依据认知工效学中的注意和记忆等机制对驾驶员的限速标志认知过程进行了解析;从驾驶员自身角度分析了影响驾驶员识别限速标志的主要因素;利用因子分析方法对调查数据和调查项目进行统计分析;最终提取了影响驾驶员识别限速标志的6个自身因素为主因子,作为限速标志对驾驶行为影响较大的因素。该研究结果可为城市道路限速标志的设计与设置提供科学依据。     

基于车桥耦合振动分析的桥梁损伤诊断方法评述

车辆通过桥梁时,桥梁和车辆的动力响应都包含桥梁结构模态或者几何参数信息,对它们进行分析能识别桥梁的模态参数和损伤。结合国内外最新研究成果,综述基于车桥耦合振动分析的桥梁结构损伤识别技术,并与传统识别方法进行比较。指出其优缺点;详细介绍基于灵敏度分析和模型修正的方法、基于结构刚度搜索的方法、利用车辆响应傅立叶变换识别桥梁频率的方法、利用车激桥梁响应的小波变换识别桥梁模态参数的方法以及综合利用车辆和桥梁响应识别桥梁损伤的方法等5种参数识别与损伤诊断方法的基本原理,并总结上述方法的实施步骤和应用时应该注意的问题;指出了该领域的关键问题和进一步的研究方向。     

隧道施工中的危险与有害因素辨识及防护

隧道工程是铁路、公路、水利和矿山建设常有的施工项目。从总体上看，随着现代科技的进步和安全管理水平的不断提高，隧道施工安全事故的发生率大大降低，但是由于种种原因，仍比其它工程作业事故的发生率相对要高。本文分析辨识了隧道施工中存在的危险危害因素，并提出了相应的安全对策与预防措施。     

化工装置火灾、爆炸、毒物扩散危险快速辨识方法

化工装置火灾、爆炸、毒物扩散的危险辨识是进一步进行定量化风险评估等安全分析的基础,笔者提出考虑该3种破坏效应显著的11种事故模式,以化工装置的介质性质、操作条件等参数作为基本事件对各种事故模式进行事故树分析;从而得到各基本参数对不同事故模式发生的影响并以其为依据建立了一种基于危险分析调查表和事故模式-字符串映射表的问卷式快速危险辨识体系;最后采用VB6.0作为开发平台,对该辨识系统进行了软件化实施。     

射频识别车辆信息管理系统

利用射频识别技术(RFID),试验开发了“射频识别车辆信息管理系统”。射频识别系统由射频卡、读取采集和系统分析与管理部分组成。系统读卡器,根据装有射频卡车辆通过感应线圈的先后顺序,判断车辆出、入场状态。通过数据采集服务程序处理存储到数据库中,供管理人员查阅分析车辆当前状态,方便了车辆调度的管理,提高了车辆的安全运行效率。该系统实用性强,管理方法科学,稳定性好。     

Statistical approach for the source identification of boron in leachates from industrial landfills

Boron has been found in high concentrations in leachates from landfills located throughout Japan. However, the source(s) of boron in the leachates, i.e., what kind of waste(s) releases this element into the leachate, has not been clarified. In this study, boron concentrations in leachates from 48 industrial landfills were evaluated, in relation to the categories of waste constituting the landfill in each of the sites, by multiple regression analysis. The multiple regression analyses were carried out using the log-transformed boron concentration as a dependent variable and each of 19 categories of industrial waste (according to the Japanese Waste Disposal and Public Cleansing Law) as independent variables. Stepwise variable selection was employed in the analyses. Although the significant variable(s) selected varied according to the data sets analyzed (viz., data sets from least controlled landfill sites, from controlled landfill sites, and from both), cinders, slag, and waste plastics emerged as wastes with positive partial regression coefficients that significantly explained the boron levels in the leachates. These results indicated that cinders, slag, and waste plastics were the sources of high concentrations of boron in the leachates. The results of the present exploratory statistical analyses warrant a systematic survey of the boron contents of, and leachability from, cinders, slag, and waste plastics. Received: January 17, 2000 / Accepted: July 24, 2000     

室内空气中多环芳烃(PAHs)的源及贡献率

采用因子分析及多元线性回归对杭州市室内外空气中PAHs的来源及其贡献率进行研究,求得PAHs总量与公共因子间的特征方程.结果表明,影响室内外空气中PAHs浓度的主要因子依次为烹调、卫生球的挥发、吸烟及加热、汽车尾气.在吸烟者家庭中,室内吸烟是影响室内空气中PAHs浓度的最重要的影响因子,其对室内空气中苯并(a)芘BaP的贡献率为25.8%.     

似然度函数对GLUE方法的影响

GLUE方法是不确定性条件下参数识别的重要方法,似然度函数的选择是GLUE方法的关键. 为克服观测误差和模型结构误差对研究结果的影响,利用合成的数据序列,以污染物衰减模型为基础,分析了似然度函数对参数识别和模型预测结果的影响. 从识别结果来看,由于似然度函数的变化改变了参数似然度之间的对比,因此对参数识别及灵敏度分析的结果都产生了影响,甚至能够改变参数全局灵敏度的相对排序. 从预测结果来看,似然度函数的变化影响了模型预测的分布,但即使预测结果的模糊性在减少,精度并没有相应地提高. 因此,似然度函数的恰当选择对分析结果具有重要的影响,应结合研究问题的具体特点和对模拟结果的要求,采用概率论等相关方法选择出尽可能体现参数真实重要性的函数.      

The different paths to post‐merger identification for employees from high and low status pre‐merger organizations

A well‐known downside of organizational mergers is that employees fail to identify with the newly formed organization. We argue that developing an understanding of factors that affect post‐merger identification requires taking the pre‐merger status of the merger partners relative to each other into account. This is because relative pre‐merger status determines employees' susceptibility to different aspects of the merger process. Specifically, for employees of a high status pre‐merger organization, we expected post‐merger identification to be strongly influenced by (i) pre‐merger identification and (ii) the perceived change in the status. In contrast, we expected post‐merger identification of employees of a low status pre‐merger organization to be strongly affected by the perceived justice of the merger event. Longitudinal data were obtained from a merger of two public sector organizations and the results supported our hypotheses. Our study shows that the extent to which pre‐merger identification, status change, and justice are important determinants of post‐merger identification depends on the relative pre‐merger status of the merger partners. Our discussion focuses on theoretical implications and managerial ramifications of these findings. Copyright © 2016 John Wiley & Sons, Ltd.