智能网联汽车对瓶颈交通流影响机理及协同优化控制方法研究

IntelligentAutomated-Connected Vehicle (ICV) is the development trend of future transportation. Due to the vehicle merging/diverging behaviors, bottleneck is the "throat" to the traffic system. Thus the bottleneck research is not only the challenge for traffic flow study, but also the key point influencing the gradual integration of ICV with the existing traffic system. The current traffic flow research pays more attention to ICV longitudinal behavior under homogeneous circumstance, lacking exploration of lateral interaction of vehicles and multilevel collaborative optimization control of traffic flow at bottlenecks. With the development of continuous data environment and experimental approaches, novel ideas and methods are provided for this study. In this study, we will firstly analyze the driving behavior characteristics of ICVs with different levels of automation/communication. Then ICV human-machine co-driving test are designed on the "multi-vehicle collaboration ICV experimental platform", to obtain driving behavior data and traffic flow data. Novel microscopic traffic flow simulation model with fusion of driver mental processes are built. Furthermore, we will explore the influencing mechanism of different ratio of ICVs on the traffic flow operation and negative effects (breakdown, capacity drop, oscillations and hysteresis). Finally, with the integration of system, lane, and individual ICV level optimization methods, a highly reliable and resilient optimization control method for bottleneck traffic flow and the cooperative driving control method for multi vehicles will be developed and tested. The research results will reveal the influence of ICVs on bottleneck traffic flow, promote driver's acceptance on ICV, and provide basic support for design of future-accommodating transportation systems.

对智能网联汽车（ICV）进入交通系统后的影响展开研究是一项面向未来且必要的前瞻性工作。车辆汇入/汇出干道的行为特征是交通流研究的难点，更是影响ICV与现有交通系统融合的关键。当前ICV环境下交通流研究缺乏对进出口瓶颈段车辆横向交互影响的探究，且鲜见对交通流的多层次协同优化控制。本研究基于全新的连续数据环境和实验手段，首先解析不同自动化/网联等级ICV行驶行为特征；然后利用“多车协作ICV实验平台”进行ICV人机共驾受控实验，获取驾驶行为及交通流数据，建立融合驾驶人心智过程的新微观交通流仿真模型；再进一步探究ICV混入对瓶颈交通流演化及运行负效应（失效、通行能力下降、震荡、磁滞）影响机理；最后整合系统、车道、个体ICV三个层面，提出运行高可靠、失效易恢复的瓶颈协同优化控制方法和多车协同行驶控制方法并进行仿真分析。研究成果将揭示ICV对瓶颈交通流的影响规律，为未来大规模使用ICV提供基础支撑。

由于在信息获取、感知能力、反应时间、交互行为等方面与传统人工驾驶车辆存在显著差异，智能网联汽车(ICV)在解决交通安全、道路拥堵以及改善驾乘体验上具有巨大潜力。快/高速路作为大都市交通的主骨架，瓶颈点则是其咽喉。因此，ICV对瓶颈交通流影响机理及协同优化控制是影响ICV逐步与现有交通系统融合的关键。课题针对混入一定比例ICV的瓶颈处复杂交通流环境，按照实验研究、系统建模、优化控制的思路展开研究，首先利用VISSIM和Visual Studio搭建了集成ICV驾驶决策模块、纵向控制模块、车辆模型模块的混合交通流集成仿真平台，并设计了仿真及参数校正流程。其次，基于换道强度分析和流密速关系理论，构建快速路合流区混合交通流系统动态模型，建立起快速路合流区通过量与混合交通流控制量之间的变化关系，并通过混合交通流仿真数据、快速路环形线圈实测数据验证了模型的有效性。最后，以提升合流区混合交通流运行效率为目标，提出了车辆协同策略。研究成果揭示了ICV对瓶颈交通流的影响规律，对智能网联环境下高/快速路合流区的交通管理具有一定的应用价值。依托上述研究成果，本课题课题共发表（录用）论文17篇（其中SCI/SSCI论文7篇，EI论文7篇），形成专著《车辆智能网联环境下交通流建模与仿真》初稿；申请发明专利6项，且全部专利均已授权；申请软件著作权1项,培养4名博士生和8名硕士研究生（含在读）。