BNs-OLS-SARIMA对城市短时交通流的预测*

通过在目标路口构建贝叶斯交通网（BNs）,并对与此交通网相关的交通流建立非平稳季节（SARIMA）模型,采用最小二乘法（OLS）取得相应模型的最佳权重组合,对缺失数据下的城市道路短时交通流进行预测。使用重庆市某路口的交通流数据对模型进行检测,通过多种预测指标对结果进行对比分析,结果表明BNs-OLS-SARIMA把交通流的网络结构与其周期性结合在一起,对短时交通流有良好的预测效果。     

道路交通状态预测研究

针对道路交通状态预测的研究成果,综述了基于实时数据分析的道路交通状态预测理论和方法的相关研究.从分析基于实时数据分析的道路交通预测的基本流程、特点和要求和单一断面交通状态预测和多断面交通状态预测角度回顾了道路交通状态预测的研究现状.结果表明:在不断发展完善单一断面预测方法的同时,需要发展以多个断面交通状态为研究对象的多断面预测方法,以促进我国智能交通系统的基础理论发展及其应用水平的提高.     

Effect of puerarin on the P13K pathway for glucose transportation and insulin signal transduction in adipocytes

To explore the effect of puerarin on insulin receptor (IR), insulin receptor substrate-1 (IRS-1) and protein expression of protein kinase B(PKB) in the P13K pathway of the glucose consumption, transportation and insulin signal transduction in 3T3-L1 adipocytes with insulin resistance. The insulin resistance 3T3-L1 adipocytes model was established by free fatty acid induction. The model cells were managed with puerarin in different concentrations. Glucose consumption was detected with glucose oxidase method, glucose transportation rate was determined by 2-deoxy-3H glucose ingesting method, and the IR, IRS-1 and PKB expression were determined by Western blot. Glucose consumption and transportation were significantly decreased in the model adipocytes, but increased after treated with puerarin (P<0.01). Moreover, the level of tyrosine phosphorylation of IR subunit β was higher in the puerarin treated groups, and that of IRS-1 was higher in the group treated with low dose puerarin than that in the model group. The 3T3-L1 adipocytes of insulin resistance model could be induced by free fatty acid successfully, puerarin could promote the glucose utilization in them to alleviate the insulin resistance, which may be related with the action in advancing the tyrosine phosphorylation of IR and IRS-1.     

Cemented backfill technology based on phosphorous gypsum

Physical-chemical properties of phosphorous gypsum, proportion and cemented mechanism of slurry with gypsum as aggregate were studied to remove the harms of gypsum pile, combining with difficult problems of excessive mined-out gobs, enormous ore body under roadway and low recovery ratio of Yongshaba Mine, Kaiyang Phosphor Mine Group, Guizhou Province, China. An appropriate backfill system and craftwork were designed, using shattering milling method to crush gypsum, double-axles mixing and strong activation mixing way to mix slurry, cemented slurry and mullock backfill alternately process. The results show that gypsum is fit for backfilling afterwards by adding fly ash, though it is not an ideal aggregate for fine granule and coagulate retardation. The suggested dosage (the mass ratio of cement to fly ash to gypsum) is 1:1:6–1:1:8 with mass fraction of solid materials 60%–63%. Slurry is transported in suspend state with non-plastic strength, and then in concretion state after backfilling. The application to mine shows the technology is feasible, and gypsum utilization ratio is up to 100%. Transportation and backfill effect is very good for paste-like slurry and drenching cemented slurry into mullock, and the compressive strength and recovery ratio are 2.0 MPa and 82.6%, respectively, with the maximum subsidence of surface only 1.307 mm. Furthermore, the investment of system is about 7×10⁶ yuan (RMB), only 1/10 of that of traditional paste backfill system. Foundation item: Project (2006BAB02A03) supported by the National Key Technology Research and Development Program; Project (08MX16) supported by Mittal Scientific and Technological Innovation Projects of Central South University during 2008     

交通运输业节能减排国际管理手段及启示

交通运输业的能源消费在许多国家中都占有较高比例,经济发展较快的国家更是如此。介绍了能源消费大国英国、美国、欧盟以及日本交通运输业节能减排采取的主要管理手段,包括健全的法律体系、节能技术创新、有效的税费政策等。借鉴其先进经验,为我国交通运输业节能减排管理工作提供启示。     

基于小波分析的交通参数组合预测方法

为了更加准确地预测交通参数的变化趋势,结合小波理论处理时变信息的优势,设计了一种在小波分析的基础上利用BP神经网络进行预测的组合预测方法。该方法采用小波理论与神经网络结合的策略,具有普适性,且比传统的基于小波分析的组合预测过程简单,为大运算量的实时应用提供了可能。利用实际数据对该方法进行了验证,并与对比算法进行了效果对比。结果表明,本文提出的方法预测效果较好,具有一定的实用性。     

城市主干路拥挤路段基于地点交通参数的行程速度估计

以提高基于地点交通参数估计主干路拥挤交通流平均行程速度的精度为目标,通过分析信号控制主干路的交通流特性及地点交通参数与平均行程速度之间的相关关系,以路段下游停车线前的截面平均速度为基础,提出了城市主干路拥挤交通流路段平均行程速度的3种估计模型,并采用仿真手段进行了对比分析。结果表明,所提出的全状态行程速度估计模型明显优于对比方法,所提出的两种分状态平均行程速度估计模型可以进一步改善全状态模型的估计效果,尤其是基于模比系数的分状态估计模型可显著提高拥挤条件下的行程速度估计精度。     

结合浮动车技术的SCATS自适应控制策略生成技术

针对SCATS信号控制系统存在的两个问题:相位差实时优化能力差和备选方案固化,尝试结合浮动车技术研究了SCATS自适应控制策略生成技术。通过改进地图匹配算法,提高浮动车路段行程时间估计精度,从而提高相位差实时优化能力以及模型的输入数据精度。同时根据实时的检测器数据和浮动车数据,采用模糊神经网络以及多属性决策技术自适应生成控制区域交叉口的配时参数。最后,以上海某SCATS控制区域为仿真路网,以SCATS采集的检测器数据及浮动车数据为仿真输入数据,以Paramics V6为仿真平台对本文模型进行验证,结果表明能够提高系统的控制效率。     

基于GPS数据及车辆运行特性分析的单车路段行程时间估计

针对现有的基于浮动车单车路段行程时间估计方法的不足,提出了一种基于GPS数据及车辆运行特性分析的单车路段行程时间估计方法。该方法主要进行了3个方面的改进:①对地图匹配模块中节点附近GPS点采用了相邻点联合匹配的方法,提高了匹配的效率;②针对不同情况下车辆在边界点前后运行特性的不同,对车辆通过边界点时间进行相应处理,提高了车辆通过路段边界点时间的估计精度;③对车辆个体在运行中因受到干扰而停车现象进行了分析,并对干扰停车予以有效剔除,降低了估计中车辆个体行为对估计结果准确性的影响。最后用出租车GPS实验数据进行了验证,结果证明了本文方法的有效性。     

基于LVQ神经网络的城市快速路事件自动检测算法

基于具有自组织功能的学习矢量量化(LVQ)神经网络设计了城市快速路异常事件的自动检测算法,提出分车道检测的构想。研究了原始数据筛选、输入向量模式、神经元个数及检测时段等参数的选择。基于小波分析技术对原始数据的高频噪声进行滤波,引入长车流量作为输入参数,并对比了引入前后的检测效果。选用加利福尼亚算法作为评价的参考依据,对其执行过程和门限值的选择进行了研究。