【作者】 胡彦梅；

【导师】 马天山；

【作者基本信息】 长安大学 ， 交通运输规划与管理， 2017， 博士

【摘要】 随着社会经济的发展,日益严重的交通拥堵已成为制约大中型城市发展的重要因素。交通流理论的研究有助于加深对交通系统运行规律的了解,可为交通运输规划、管理和控制提供理论指导,还可为先进的交通系统和技术的测试提供理论基础。因此,研究交通流复杂动态特性在理论和实践两方面都具有重要意义。在车联网环境下,驾驶员可以利用多个邻近车辆的运动信息,预期并调整自身车辆的速度,这种多预期的驾驶策略,可以促使车辆协同行驶在队列中,有助于缓解交通拥堵,提高道路通行能力。驾驶员的反应延迟时间是不可避免的,在交通流模型中引入驾驶员的反应延迟能够更好地描述实际的交通流特性。实际交通环境中普遍存在的是多车道、多车种构成的混合交通流,为了更加准确地刻画实际交通,需要研究多车道模型和多车种模型。考虑多预期驾驶策略、驾驶员反应延迟和多车道等因素,建立交通流模型,并对模型进行理论分析和动态特性模拟,研究多车种LWR(Lighthill-Whitham-Richards)交通流模型的高分辨率数值方法。主要研究工作包括以下几个方面:1.在车联网环境下,基于向前看驾驶策略,构建多预期延迟模型模拟交通流的复杂动态特性,并对模型进行稳定性分析。考虑非均衡地使用前方多个相邻车辆间的车头间距与速度差信息,并引入驾驶员的反应延迟时间构建了一个多预期跟驰模型。数值仿真结果表明,为尽可能少地引入输入变量,不均衡地利用前车的车间距和速度差信息是必要的。通过引入当前车辆分别与前方多辆车间的速度差信息和驾驶员反应延迟时间,提出另外一个多预期跟驰模型。数值模拟结果表明,适当地引入当前车辆分别与前方多辆车间的速度差可以增强交通流的稳定性,在一定程度上可以缓解交通拥堵。构建包含驾驶员反应延迟时间的多预期格子流体力学模型,研究结果表明延迟时间对稳定性所造成的不利影响可通过增加考虑前方的格子数量来适度弥补。文中构建的三个多预期延迟模型可为车联网环境下合作驾驶的交通流仿真提供基础模型。2.在车联网环境下,基于双向视野的驾驶策略,构建一个多预期跟驰模型,再将其推广为宏观模型,应用理论分析和数值模拟方法分别研究新宏观模型的稳定性条件和交通流的局部集簇效应。数值模拟与稳定性分析的结果相符合,验证了理论分析的合理性。构建的双向视野的宏观模型可用于车联网环境下基础设施的新建、扩建及宏观管理措施制定时对交通流的宏观模拟。3.建立一种扩展的流量转移率描述车辆的换道行为,构建新的三车道格子流体力学模型,并应用线性稳定性方法推导模型的稳定性条件。通过研究小扰动情形下车辆的换道行为,证实了扩展的流量转移率和改进模型的合理性。新的三车道交通流格子流体力学模型可为混合交通流中车辆换道的交通流仿真提供基础模型。4.将无振荡中心格式推广于多车种LWR交通流模型,给出一种求解模型的高分辨率数值方法。运用非线性限制器近似离散斜率和通量的空间导数,保证格式无振荡且不需计算雅可比矩阵。方法形式简单、计算量小,同时又具有较高的分辨率。数值算例的结果表明了方法的有效性。更多还原

【Abstract】 With the rapid development of economy,the increasingly serious traffic congestion has become an important factor restricting the development of large and medium-sized cities.The research of traffic flow models and theory can not only help to deepen the understanding of the law of traffic system operation,but also act as theoretical guidance for transportation planning management and control,and moreover,can provide theoretical basis for the test of advanced transportation system and technology.Therefore it is of great significance in theory and practice to study the modelling of complex dynamic properties of traffic flow.In a connected vehicle environment,a driver can use the movement information of some adjacent vehicles to anticipate and adjust the velocity of his own vehicle.By using such multi-anticipation driving strategy,every vehicle in the system can cooperate with its adjacent vehicles to run in a queue,which helps to ease traffic congestion and increase road capacity.The reaction-time delay of drivers is unavoidable.Therefore the introduction of the reaction-time delay of drivers in the traffic flow models can better describe the real traffic flow characteristics.In the actual traffic,it is common that the traffic is mixed by multi-lane road and multi-class vehicles.Thus,the research of multi-lane models and multi-class models is necessary in order to describe the actual traffic flow more accurately.Several novel traffic flow models are constructed by considering the multi-anticipation driving strategy,reaction-time delay of drivers or multi-lane traffic,and the proposed models are investigated based on the theoretical analysis and the simulations of dynamic properties.Moreover,a high-resolution numerical method for solving the multi-class LWR(Lighthill-Whitham-Richards)model is investigated.The main work is as follows:1.Based on the forward-looking driving strategy,the multi-anticipation delayed models are constructed to simulate the complex dynamic characteristics of the traffic flow in a connected vehicle environment,and the stability analyses are conducted for the novel models.A multi-anticipation delayed car-following model is constructed by introducing the reaction-time delay of drivers and considering the use of the unbalanced information about headways and velocity differences between multiple adjacent vehicles in front of the current vehicle.The results of numerical simulations demonstrate that it is necessary to unevenly use information of headways and velocity differences of front vehicles in order to introduce variables as few as possible.Another multi-anticipation car-following model is proposed by introducing the reaction-time delay of drivers and the multiple velocity differences between the current vehicle and the preceeding vehicles.The numerical results show that appropriate introduction of multiple velocity differences can enhance the stability of traffic flow,and can ease the traffic congestion to a certain extent.A multi-anticipation lattice hydrodynamic model is presented by taking into account drivers’ reaction-time delay and multi-anticipation driving strategy.The simulating results show that the unfavorable impact caused by the reaction-time delay on the stability can be partly compensated by introducing more lattices in the model.The three constructed multi-anticipation delayed models can provide the basic models for the traffic flow simulation of cooperative driving in the network environment.2.In the connected vehicle environment,a multi-anticipative car-flowing model is constructed based on the driving strategy of the bidirectional perspective.The obtained model is extended to a macro model.The stability condition and local cluster effect are investigated for the macro model via the theoretical and numerical methods,respectively.The simulation results are consistent with the results of the stability analysis,which confirms the rationality of theoretical analysis.The novel macro model of the bidirectional perspective can be used as the basic model to simulate the traffic flow by the macro way in the connected vehicle environment,which helps to study the construction and expansion of infrastructure and formulate macro management measures.3.An extended transfer flow rate is proposed to mimic the lane-changing behavior.A new three-lane lattice hydrodynamic model is presented.The stability condition of the model is derived by applying the linear stability method.For the case of small disturbance,the behavior of the lane changing is investigated.The results show that the extended transfer rate and the improved model are rationality.The new three-lane lattice hydrodynamic model can be a basic model for simulation traffic flow to investigate the lane-changing behavior in mixed traffic.4.The non-oscillatory central scheme is extended to the multi-class LWR traffic flow model.A high-resolution numerical method for solving the model is presented.The discrete slopes and the spatial derivative of the flux are approximated by a nonlinear limiter to guarantee the non-oscillatory behavior and to avoid the computation of the Jacobian.The resulting method has the advantages of simple form,small computation and high resolution.The results of numerical experiments demonstrate the efficiency of the method.更多还原