多方式多类型交通分配模型(MMA)在收费型公路网中的应用研究

在对收费型公路运营模式进行研究的基础上,利用交通规划软件建立了基于广义费用的交通流量分配模型即多方式多类型交通分配模型,通过对模型参数的科学标定和网络的合理设置,得到各类车型在特定网络上的流量分配、路段饱和率以及通行费用,并对此模型的应用效果进行分析,最终为区域高速公路网规划、路段交通量预测提供依据。     

Vehicle sensor data-based analysis on the driving style differences between operating indoor simulator and on-road instrumented vehicle

Abstract

Indoor simulator and on-road instrumented vehicle are the most popular ways to analyze driving behaviors by using collected Vehicle Sensor Data (VSD). However, for a same driver, the driving performance could be different in the real world and in the simulated world. Even though many studies have been conducted to discover the differences of driving behaviors in these two circumstances, little research has focused on analyzing the differences in driving style, which can provide more integrated knowledge of a driver from the natural structure, stimulus–response mechanism, of driving behaviors. Therefore, in this paper, the driving styles in both the real world and the simulated world are extracted by implementing the nonnegative matrix factorization method on the collected VSD data. Through this analysis, the driving style differences can be quantitatively described and discussed in detail. It is found that the drivers tend to be more unstable and sometimes aggressive when driving the simulator and the deviation in the perception of temporal gap in two circumstances is also discovered. The research findings are particularly valuable to calibrate the driving simulator and construct more reliable driving behavior models.     

考虑沥青路面材料参数空间差异性的解析计算及影响分析

复杂服役环境下沥青路面的材料参数存在明显的空间差异性,具体表现为沿深度的模量梯度分布与非连续层间接触以及横观各向同性,实现其解析计算对现有的沥青路面力学响应分析及结构设计方法具有重要的参考价值。为此,从弹性力学基本方程出发,基于状态空间法,推导了具有指数模量梯度的横观各向同性弹性层状体系通解;在此基础上,基于波传递方法,将状态向量转换为波向量,并结合边界条件和层间条件,建立了考虑沥青路面材料参数差异性的多层路面结构解析解;然后借助MATLAB平台编制了数值计算程序,采用文献对比、BISAR程序以及有限元软件ABAQUS验证了该解析解的精度与效率;最后结合典型数值算例,阐述了该解析解在沥青Top-Down开裂和传统路面疲劳损伤分析中的应用。理论推导和数值计算结果表明：该解析解摆脱了弹性力学问题求解对应力函数的依赖性,并且借助波传播方法,避免了求解中正指数过大对求解精度的影响,并使得边界条件和不同层间接触关系的描述更加便捷;沥青路面材料参数的空间差异性不可忽视,沥青面层连续模量梯度会造成更大的路表拉应力和最大剪应力,进而加剧沥青路面的Top-Down开裂;而横观各向同性和层间接触状态间存在耦合效应,从而会加剧沥青路面的疲劳损伤。     

兰州地铁地温分布特性及其演化规律

以兰州地铁所在地区为研究对象,实测地铁隧道开挖前的地温(简称为初始地温),根据实测数据,提出地铁初始地温预测模型公式。采用非稳态传热的数值模型,分析运营条件下地铁隧道围岩温度的演化规律。结果表明:兰州地铁初始地温随环境气温和埋深的变化而变化;年变温层位于自地表至埋深12m处;年恒温层位于埋深12m及其以下,温度为15℃左右;年变温层中,1年内初始地温变化规律与环境气温变化规律相似,近似呈正弦曲线状分布,但存在相位滞后的现象;1年中初始地温的振幅随埋深的增大呈指数下降趋势。在隧道内空气与围岩之间热交换中,兰州地铁隧道围岩的温度及其梯度、热透厚度(未达到极限时)均与隧道内环境温度、热交换时间成正相关关系,但与距隧道内壁的距离成负相关关系。     

灰铸铁飞轮疲劳失效分析

大马力飞轮是商用车一款大马力发动机的零部件,其搭载大马力发动机为商用车提供动力总成。飞轮在工作期间,因经常与离合器摩擦,售后市场出现离合器打滑,起步困难、抖动较大等整车问题,拆解后发现,飞轮表面出现磨损不平、表面微裂纹为主要的失效形态的现象。     

基于分布质量轴模型的回旋振动计算方法

根据旋转Timoshenko梁的运动微分方程,推导了计算船舶推进轴系回旋振动时分布质量轴模型的传递矩阵。相对于集中质量轴模型,分布质量轴模型可以计入轴分布质量的平动惯量、转动惯量、陀螺力矩以及剪切变形的影响,更加逼近船舶推进轴系回旋运动的实际物理模型,因此能更精确地计算轴系的临界转速和固有振型。在分布质量轴模型的基础上,采用数值稳定性良好的改进Riccati传递矩阵法编制了回旋振动计算程序,并应用于两型船舶的推进轴系计算。将文中方法的计算结果与有限元法以及第三方计算报告进行了比对,结果具有良好的一致性,从而验证了方法的正确性和计算精度。     

盲源分离技术在水声信号中的应用研究

通过研究α稳定分布,提出海洋噪声、舰船辐射噪声等符合低阶α稳定分布,在进行对称α稳定分布且独立的随机源信号进行盲分离处理时,若仍然采用高斯分布将不收敛。本文设计了分数低阶矩阵盲源分离算法,最后通过实验对比本文所设计算法和时延分离算法。实验结果表明:本文所设计的算法对于α稳定分布且有独立的随机信号,具有较好的收敛性,能够有效实现盲源信号分离。通过从复杂的海洋噪声中仍能较好地提取出船舱辐射噪声的实验,说明即使信噪比发生变化,源信号分离后相似系数较分离前仍然有稳定地提高。     

张力缆系泊浮体的数值波浪水槽时域模拟

为了能够快速真实地模拟波浪与张力缆系泊浮体相互作用问题,基于二维边界元格式建立了线性条件下的数值波浪水槽,并以此为基础对张力缆系泊浮体与波浪相互作用问题进行了时域数值模拟。张力缆的作用以系泊刚度的形式表达。与其他学者的相关研究结果对比表明该计算程序是可靠的,该方法可以为相关海洋工程的设计研究提供有利的参考。     

流场中变截面加筋柱壳结构声振特性分析

基于传递矩阵法给出了流场中变截面加筋柱壳结构在受集中力和声压作用下振动与声辐射的求解过程,并在确定模态截断算法有效性的基础上开展对结构振动特性的研究,分析了环板-圆柱壳加筋柱壳结构的损耗因子、流体介质、壳体厚度、环肋数目对其声振特性的影响.结果表明,流体介质的存在使结构振动减小,辐射声压增加;损耗因子增加,振动和辐射声压在中高频段降低;壳体厚度增大,壳体的振动响应减小,低频段辐射声压降低,高频段辐射声压交叉波动;环肋数目增加,结构振动减小,在中高频段,辐射声压降低.     

Human-like autonomous car-following model with deep reinforcement learning

This study proposes a framework for human-like autonomous car-following planning based on deep reinforcement learning (deep RL). Historical driving data are fed into a simulation environment where an RL agent learns from trial and error interactions based on a reward function that signals how much the agent deviates from the empirical data. Through these interactions, an optimal policy, or car-following model that maps in a human-like way from speed, relative speed between a lead and following vehicle, and inter-vehicle spacing to acceleration of a following vehicle is finally obtained. The model can be continuously updated when more data are fed in. Two thousand car-following periods extracted from the 2015 Shanghai Naturalistic Driving Study were used to train the model and compare its performance with that of traditional and recent data-driven car-following models. As shown by this study’s results, a deep deterministic policy gradient car-following model that uses disparity between simulated and observed speed as the reward function and considers a reaction delay of 1 s, denoted as DDPGvRT, can reproduce human-like car-following behavior with higher accuracy than traditional and recent data-driven car-following models. Specifically, the DDPGvRT model has a spacing validation error of 18% and speed validation error of 5%, which are less than those of other models, including the intelligent driver model, models based on locally weighted regression, and conventional neural network-based models. Moreover, the DDPGvRT demonstrates good capability of generalization to various driving situations and can adapt to different drivers by continuously learning. This study demonstrates that reinforcement learning methodology can offer insight into driver behavior and can contribute to the development of human-like autonomous driving algorithms and traffic-flow models.