收敛系统的性质

离散时间收敛系统是研究离散时间线性模型渐近理论的重要工具。利用随机积分工具推广了离散时间收敛系统的概念,给出收敛系统的定义并获得了收敛系统的几个重要性质。     

Understanding cyclists’ perceptions,keys for a successful bicycle promotion

Many variables that influence bicycle use beyond time and cost have been included in models of various types. However, psycho-social factors that make the bicycle eligible as a modal alternative have not been identified properly. These factors are related to intention, attitudes and perceptions, and their identification can contribute to obtain the keys for a successful bicycle policy. Here, an in-depth investigation of cyclists’ perceptions is attempted using a large university survey designed and collected ad hoc, and then applying exploratory and confirmatory factor analyses. After identifying fourteen factors, a structural equations model was estimated to find structure and relationships among variables and to understand users’ intentions to use the bike. Four (latent) variables are identified, namely convenience, pro-bike, physical determinants and exogenous restrictions. The main conclusion is that convenience (flexible, efficient) and exogenous restrictions (danger, vandalism, facilities) are the most important elements to understand the attitudes towards the bicycle.     

A methodology for classification by priority for action: Selecting road stretches for network noise action plans

Road traffic noise is an element of outstanding importance within the overall context of environmental impact. This problem must be technically addressed from an efficient viewpoint, and solutions or alternatives should be considered by means of appropriate and consolidated procedures.Up to now, there is no regulated guideline for establishing well-founded priorities when dealing with the diverse road stretches included in the corresponding Action Plans against noise under the Environmental Noise Directive (2002/49/EC). To this end, the present study proposes a methodology to sort, by priority, road stretches identified by their noise problems and therefore requiring appropriate action. The methodology is based on the so-called “road stretch priority index” (henceforth referred to as RSPI). This index involves a number of variables (called “road stretch priority variables”) that are weighted according to their influence on the road traffic noise problem. Thus, the RSPI makes it possible to prioritize different stretches of the Action Plan. To illustrate the application of the proposed methodology, this paper also describes a real case entailing a difficult choice, applying the proposed methodology to a review of the Action Plan against Noise 2008–2012 in the province of Almería, for the road network of the regional government of Andalusia, Spain.     

多变量聚类分析的高速公路交通流状态实时评估

针对交通状态单变量判断传统方法的不足,本文建立了基于多变量聚类分析的高速公路交通流状态实时评估方法.结合实际交通流数据,利用模糊聚类、K均值聚类等算法对速度、流量等向量进行聚类分析,给出适合当前高速公路特点的交通状况划分方法和关键参数.本文方法能够实时、准确、全面地反映交通流的运行情况,为制定高效的交通管理控制方案及合理的出行方案提供数据基础.     

基于集成模型的多维活动-出行决策研究

对个体活动-出行行为的5个决策维度进行联合建模分析,包括连续型的通勤距离、离散型的日通勤出行方式和通勤出发时间、小汽车拥有和日非工作活动量.一方面,作为内生变量,这5个决策维度受到个体性别、收入、教育水平等社会经济属性影响;另一方面,它们之间的相互影响关系也通过可见的结构关系及不可见的相关关系得到证实.比如,通勤出行距离越长,出行者家庭小汽车的拥有量越大;通勤出行距离、公交出行方式的选择均对日非工作活动量有负面影响;当通勤者家庭拥有更多的小汽车或者通勤出行距离很长时,他们选择小汽车和公交车的概率更高.本文证实了个体活动-出行决策维度之间的复杂作用关系,证实了对多维活动-出行决策进行综合集成分析的重要性.     

相关跳频系统中一种高效FFT变换算法

相关跳频是一种先进的扩频技术，具有跳速快、实现简单、抗干扰性能力强等优点。FFT变换算法是相关跳频的重要技术。对FFT变换算法进行了研究并提出了一种改进，减少了运算次数，从而大大加快了运算速度。该算法可以用于相关跳频系统中。     

基于变点分析思想的高维机械噪声数据本征维估计方法

利用变点分析思想,针对高维数据协方差矩阵的特征值曲线特点,提出一种本征维估计方法:基于均值估计的特征值变点法,并应用于某双层圆柱壳体的机械噪声数据处理中。通过与其他现有常用方法的比较分析,验证了基于均值估计特征值变点法的可行性。结果显示:基于均值估计的特征值变点法理论简单、计算简便,计算结果客观、合理,对某双层圆柱壳体机械噪声数据的处理效果较好。     

线程同步中的条件变量控制应用

针对多线程并发中的竞争和协作控制问题,介绍了Linux环境下的条件变量机制及其相关系统调用,并对其工作原理进行分析.以线程同步中的司机、售票员模型为例,指出一般信号量解决方案的不足,提出基于条件变量机制的解决方案,并给出一个具体的控制模型.     

Hilbert空间映射下非线性分类的变元可分离核函数

分析变元可分离函数K（u,v）=k（u）k（v）成为核函数对非线性支持向量分类机的作用,由新方法构造的非线性支持向量分类机可以使相应的算法得以简化.     

基于无记忆非线性变换的相关非高斯雷达杂波的仿真

大量的实测数据表明,雷达杂波的分布特性明显偏离高斯分布。相关非高斯杂波的仿真主要包括对数正态、韦伯尔、K分布杂波的仿真。在杂波信号模拟中,需同时考虑杂波幅度概率分布和其相关特性。文中利用广义维纳过程的无记忆非线性变换法（ZMNL）来建模仿真,产生相关非高斯杂波。该仿真系统不仅可以实时动态地模拟实际采集的雷达杂波,而且可按给定的功率谱模型来产生满足非高斯分布的随机数据序列。仿真结果证明了此方法的有效性。