Examining driver behavior using data gathered from red light photo enforcement cameras

INTRODUCTION: To develop an effective driver assistance system that can prevent red light violations, it is important to have a thorough understanding of contributing factors and circumstances surrounding such violations. METHODS: Red light violation behavior was examined using about 47,000 violation records that were captured by photo enforcement cameras from 11 signalized intersections in the city of Sacramento, California, over a four-year period. This examination identified factors with strong correlation to red light running behavior. RESULTS: Key findings based on Sacramento's red light violation records include: (a) younger drivers showed a higher tendency of running the red light and were more likely to commit such a violation at speeds above the posted speed limit; (b) off-peak time period between 8 p.m. to 5 a.m. had lower violation counts, but red light violators in this time frame had a higher propensity of racing through intersections at high speeds; (c) also during the time period between 8 p.m. to 5 a.m., violators showed a higher probability of entering intersections two or more seconds after the onset of red light; and (d) violators were less likely to cross high traffic-volume intersections with speeds greater than the posted speed limit. DISCUSSION: Sacramento's data suggest that some red light violators might have committed such an offense unintentionally. The top three violation speeds were all below 20 mph and more than 94% of red light runners crossed intersections within 2 seconds after the onset of the red light. IMPACT ON INDUSTRY: In the next several years, studies will be executed to assess whether a red light violation warning system can reduce "unintentional red light running" such as drivers who made incorrect judgment regarding the adequate time to clear the intersection and inattentive motorists who fail to observe the presence and status of the red light.     

基于自适应网络模糊推理方法在河道洪水预报中的应用

对于易受洪灾的地区而言,快速而准确的洪水预报非常重要,能够为洪水预警消息的发布提供更长的先导时间,从而为可能受灾地区的人们提供更充足的时间以采取相应的防洪措施或安全转移。 常用的预报模型包括基于物理性模型和基于系统技术模型。尽管物理性模型能对洪水形成的物理过程提供很好的解释, 水文学家并不愿意使用它们,因为模型中参数的率定是比较复杂的。因此,一种基于纯数据集的黑箱技术已被广泛采纳。常用的黑箱模型包括线性模型(LR)、自回归移动平均模型(ARMA)和人工神经网络模型（ANN）等。 在当前的研究中,一个相对新颖的黑箱模型--基于自适应网络的模糊推理系统（ANFIS）被用来对长江某河段的洪水进行预报。与此同时,一个线性回归模型（LR）用来作为ANFIS模型的对照。在构建ANFIS中,混合学习算法 (即误差反衍（BP）耦合最小二乘法（LSE）) 用来训练模型的参数。此外,为避免出现过度训练现象,原始数据集基于统计特征值划分成3个子集：训练集、测试集和校正集。当对ANFIS模型训练时,测试集用来帮助控制训练代数。结果表明,ANFIS的预报效果优于LR模型。分析认为ANFIS能够提供预报精度是因为其采用了局部拟合技术,通常它会优于LR模型所采用的全局拟合技术。最后,对本研究而言,最适合的ANFIS模型是输入量为梯形的成员度函数。     

体系安全性分析的双层次框架

为识别体系(So S)中的薄弱环节、评价威胁风险并提高体系安全性,根据安全系统工程和体系工程理论,提出体系安全性基本概念。基于复杂网络科学和传统安全分析技术,提出一种双层次分析框架,从脆弱性角度和威胁风险角度研究体系安全性。脆弱性分析可从拓扑结构和非拓扑因素2方面识别对体系安全较关键的脆弱点;威胁风险分析能得到脆弱点面临的任意威胁组合,并识别后果较严重的威胁风险。结果表明:体系安全性包括宏观的脆弱性应对能力,和微观的威胁风险控制能力。体系安全性分析需宏观和微观双层次相结合,全面考虑体系安全性诸因素。基于识别结果的威胁评价将提高安全分析的针对性和效率。     

Generalized analytical expressions for safety instrumented systems' performance measures: PFDavg and PFH

Safety Instrumented Systems (SIS) constitute an indispensable element in the process of risk reduction for almost all of nowadays' industrial facilities. The main purpose of this paper is to develop a set of generalized and simplified analytical expressions for two commonly employed metrics to assess the performance of SIS in terms of safety integrity, namely: the Average Probability of Failure on Demand (PFD_avg) and the Probability of Dangerous Failure per Hour (PFH). In addition to the capability to treat any K-out-of-N architecture, the proposed formulas can smoothly take into account the contributions of Partial Stroke Testing (PST) and Common Cause Failures (CCF). The validity of the suggested analytical expressions is ensured through various comparisons that are carried out at different stages of their construction.     

Heuristics as adaptive decision strategies in management

In the management literature, heuristics are often conceived of as a source of systematic error, whereas logic and statistics are regarded as the sine qua non of good decision making. Yet, this view can be incorrect for decisions made under uncertainty, as opposed to risk. Research on fast and frugal heuristics shows that simple heuristics can be successful in complex, uncertain environments and also when and why this is the case. This article describes the conceptual framework of heuristics as adaptive decision strategies and connects it with the managerial literature. We review five classes of heuristics, analyze their common building blocks, and show how these are applied in managerial decision making. We conclude by highlighting some prominent opportunities for future research in the field. In the uncertain world of management, simple heuristics can lead to better and faster decisions than complex statistical procedures. Copyright © 2014 John Wiley & Sons, Ltd.     

Local food networks as catalysts for food policy change to improve health and build the economy

Local food networks (LFNs) are growing in popularity, in part as a response to broader criticisms of conventional food production. Municipal policy-makers have the opportunity to work with stakeholders to build LFNs to increase access to healthy foods in cities and ultimately improve population health and well-being. Building opportunities for healthy eating is particularly important in our study area. Flint, Michigan, is a post-industrial shrinking city suffering from the economic and health effects of deindustrialisation. Various stakeholders in Flint have responded to a significant issue with access to food by strengthening collaborations through a food policy council (FPC). Growth in the local food system has been supported by administrators and community advocates alike, through supporting community gardens, farmers' markets, and urban agriculture in a manner similar to nearby Detroit. Participant observation was conducted with stakeholders involved in the development of the LFN and the FPC in Flint. Stakeholders were exposed to existing research on the food system to help inform their policy direction. The group expressed several core concerns and prospects for future work, including a strong emphasis on consensus-based decision-making. Based on the synthesis of stakeholder opinions, policy recommendations are made to aid in continued planning of the LFN. Planning for food is an important first step in improving public health and strengthening local economic development in post-industrial cities. This research highlights the issue by making explicit the challenges and opportunities for policy advocacy in LFNs.     

Methane and nitrous oxide emissions from a subtropical coastal embayment(Moreton Bay,Australia)

Surface water methane(CH4) and nitrous oxide(N2O) concentrations and fluxes were investigated in two subtropical coastal embayments(Bramble Bay and Deception Bay,which are part of the greater Moreton Bay, Australia). Measurements were done at 23 stations in seven campaigns covering different seasons during 2010–2012. Water–air fluxes were estimated using the Thin Boundary Layer approach with a combination of wind and currents-based models for the estimation of the gas transfer velocities. The two bays were strong sources of both CH4 and N2O with no significant differences in the degree of saturation of both gases between them during all measurement campaigns. Both CH4 and N2O concentrations had strong temporal but minimal spatial variability in both bays.During the seven seasons, CH4 varied between 500% and 4000% saturation while N2O varied between 128 and 255% in the two bays. Average seasonal CH4 fluxes for the two bays varied between 0.5 ± 0.2 and 6.0 ± 1.5 mg CH24/(m·day) while N2 O varied between 0.4 ± 0.1 and1.6 ± 0.6 mg N2O/(m2·day). Weighted emissions(t CO2-e) were 63%–90% N2 O dominated implying that a reduction in N2 O inputs and/or nitrogen availability in the bays may significantly reduce the bays' greenhouse gas(GHG) budget. Emissions data for tropical and subtropical systems is still scarce. This work found subtropical bays to be significant aquatic sources of both CH4 and N2O and puts the estimated fluxes into the global context with measurements done from other climatic regions.     

Investigation of drinking water bacterial community through high-throughput sequencing

<正>Delivery of safe and pathogen-free drinking water is crucial to public health.However,there exist challenges to the maintenance of the sterility of drinking water throughout the drinking water distribution systems(DWDS).Microbial growth in DWDS,such as growth of opportunistic pathogenic microorganisms,can lead to severe health problems in consumers(Berry et al.,2006;Brettar and Hofle,2006;Lu et al.,2014;Zhang et al.,2015).     

An intelligent disaster decision support system for increasing the sustainability of transport networks

The increase in extreme weather events arising from climate change is posing serious threats to the sustainability of transport systems, creating the need for improved tools for decision support for more effectively managing natural disasters. There are numerous transport‐related decisions that are required during the response, recovery and preparedness stages of the disaster management cycle. This paper describes the development and application of the Intelligent Disaster Decision Support System (IDDSS), which provides a platform for integrating a vast range of road network, traffic, geographic, economic and meteorological data, as well as dynamic disaster and transport models. Initial applications to the response and planning for floods and fires are presented to illustrate some of its capabilities. The IDDSS can be used to improve disaster management, which in turn will increase the sustainability of transport networks.     

Analysis on causes of chemical industry accident from 2015 to 2020 in Chinese mainland: A complex network theory approach

Chemical manufacturing is a long-process industry, where an end product may pass through numerous dangerous and complex steps. In such long chains of coordinated activity, accidents remain common. This study made loss-prevention recommendations for the chemical industry after conducting a review of accident reports and creating a complex network model. A human factor analysis and classification system (HFACS) was used to classify data from 109 investigation reports from the Chinese mainland (2015–2020). Levels Ⅱ and Ⅲ of the HFACS output were fed into a complex network model to generate a map of causes and chains of risk. It was shown that most accidents were directly or indirectly caused by human action, and human factors played a decisive role in occurrence, evolution, and resolution. The model used was visualized in Gephi, and the key cause nodes were identified by their topological characteristics. A modularity algorithm was used to derive the community structures and segment the network map. Crucial nodes in each community were compared with factors for each class in the HFACS model. It was also found that there was a biasing factor in the causal processes of explosive accidents and poisoning and asphyxiation accidents according to the associations classified by modularity. Risk abatement strategies were proposed for the crucial factors.