基于粗集和神经网络耦合的短时交通流预测

比较分析神经网络和粗糙集在数据处理过程中的各自优缺点,提出一种基于二者强耦合集成方式的短时交通流预测模型。首先利用粗集对获取的交通流数据进行预处理,简化神经网络训练样本数据集并通过粗集属性约简提取决策规则;其次,利用所提取的规则直接确定神经网络的隐层数、隐层节点数及节点的相互关系;最后训练神经网络用于短时交通流预测。通过与单纯利用神经网络预测的结果进行比较,发现该模型降低了网络训练时间,提高了预测精度。     

����·�������г�ʱ��Ԥ���о�

本文提出了一种基于状态空间神经网络（SSNN）和拓展卡尔曼滤波（EKF）的混合式行程时间预测模型。以往的研究表明，状态空间神经网络能够较好的处理复杂的非线性时空问题。然而，状态空间神经网络需要大量的历史数据作为离线训练之用。其不足之处在于，首先是需要花费大量的时间和精力去收集、准备数据，以及训练神经网络。其次，输入输出随着时间不断增加，训练过程需要不断的从新重复。为了提高状态空间神经网络的有效性，扩展卡尔曼滤波代替了传统的方法来对神经网络进行训练。荷兰的一条城市道路被选择为模型验证的试验路段。通过与另外两个预测模型之间的对比验证，该模型的预测能力能够达到满意的有效性、准确性和鲁棒性。     

Agricultural vehicles and sustainable safe road traffic: solving conflicts on arterial highways

Addressing the issues of traffic safety in rural areas presents a constant challenge. The mix of light and heavy vehicles and the considerable differences in speed among these traffic participants result in high risks and delays for the faster vehicles. Agricultural vehicles (AVs) in particular have such an impact on traffic, especially when using arterial highways. This paper reviews the problems of safety and delays that AVs cause on arterial highways, and the appropriate mitigation. The concept of 'sustainable safety' in The Netherlands focuses on these problems, because of the proposed construction of parallel roads alongside all arterial highways. However, Dutch accident statistics cannot justify the high costs for the construction of parallel roads alongside 7000 km of arterial highways. Delays experienced by fast traffic are another reason for separating AVs from other road users with parallel roads. Alternative measures alongside the arterial highway, such as passing bays, restricting AVs to travelling at off-peak only and improving the conspicuity of the AVs, may be more cost-effective ways of reducing delays and/or improving traffic safety on arterial highways. Another solution may be to eliminate the need for AVs to use the arterial highway by altering their routes. For this purpose, land reallocation projects (as practised in Holland) can provide a useful tool.     

Advanced SiC Power Modules for 13.8 kV Power Distribution

Northrop Grumman Corporation has been developing 10 kV SiC MOSFETs and Junction Barrier Schottky diodes for application to a 13.8 kV 2.7 MVA solid-state power substation. The design of half-bridge power modules has extensively used simulation, from electron-level device simulations to the system-level trade studies, to develop the most efficient module for use in the SSPS.     

峡湾与福特GT 挪威远行之旅

正似乎没花什么功夫全新福特GT就让大家承认了它是一台从头纯粹到尾的超级跑车。但名字叫做GT就要有实际的公路行驶能力,于是我们选了一台亮黄色的福特GT前往挪威,去看看它在著名挪威大西洋公路和北极圈赛道Arctic Circle Raceway上的表现能否令人信服有些时候选题会出乎意料地顺利,提议受到欣赏,计划顺利执行,需要的酒店机票租车统统很快订好;但多数时候我们都不得不再三更改计划,甚至干脆推倒重来。当然,最终的结果基本还是能让人满意,偶尔还会超过大家的预期。这一次便如此,毕竟在挪威蓝色月光     

�г�ʱ���������ݴ������о�

车牌照匹配技术的广泛应用为行程时间样本数据的精确获取提供了一种可行的方式。但是,由于各种客观原因,采集到的原始数据夹杂着大量的噪声数据,只有剔除了这些噪声数据,行程时间样本数据才能更好地应用于行程时间分析与服务。本文对产生噪声数据的原因进行了详细的分析,回顾了现有的噪声剔除方法,并分析其可能存在的问题;基于交通信息提取计算模型,提出了一种噪声剔除的新方法。以荷兰代尔夫特的一条城市道路为研究对象,基于实际数据对新方法和现有方法进行了对比分析。结果表明,新方法能够更好地剔除噪声数据。