This paper focuses on two fundamental issues in traffic flow modelling: string stability of car following (CF), and oscillation of traffic flow. Its aim is to explore the complementary use of CF instability analysis and oscillation analysis to compress and ease traffic congestion in a connected environment. Each of these topics has been extensively investigated in the literature. However, each topic has been investigated separately and, despite the inherent conceptual and empirical closeness of the two concepts, little effort has been devoted to untangling their relationship.

To address this failing, we first define four types of oscillation – amplitude-decay oscillation, amplitude-ceiling oscillation, speed-deviation ceiling oscillation, and speed-deviation growth oscillation – and reveal their similarities and dissimilarities to CF instability. Based on the stability criterion, we then develop oscillation criteria to identify different types of oscillation by relaxing two unrealistic assumptions used in CF stability analysis (i.e., infinitely-long platoon and long-wavelength perturbation). Finally, to demonstrate how CF instability analysis and oscillation analysis can be combined to influence individual vehicle stability and improve traffic oscillations in a connected environment, a platoon of vehicles that experience an oscillation in the NGSIM data is used in a case study. In this case study, different control factors are used for different vehicle types: connected (but human-driven) vehicles, and automated vehicles.

Our analysis shows that a higher stability of some individual vehicles can alleviate the oscillation severity for the platoon. It also shows that desired time gap and maximum acceleration are two promising parameters that can be used to both improve individual vehicle stability and significantly smooth the oscillation of the platoon in a connected and/or automated environment. Of particular note, when considering all of the factors explored in our analysis, adjustment of the desired time gap is the most effective factor in smoothing traffic oscillations.

本文着重研究交通流建模中的两个基本问题：汽车跟随（CF）的弦稳定性和交通流的振荡。其目的是探索CF不稳定性分析和振荡分析的补充使用，以压缩和缓解连接环境中的交通拥堵。这些主题中的每一个都在文献中得到了广泛的研究。但是，每个主题都经过单独研究，尽管这两个概念在概念上和经验上具有内在的亲密性，但很少花精力来理清它们之间的关系。

为了解决这种故障，我们首先定义了四种类型的振荡-幅值衰减振荡，幅值上限振荡，速度偏差上限振荡和速度偏差增长振荡-并揭示了它们与CF不稳定性的异同。基于稳定性准则，我们然后通过放宽CF稳定性分析中使用的两个不切实际的假设（即无限长排和长波长扰动）来开发振荡准则，以识别不同类型的振荡。最后，为了说明CF不稳定性分析和振动分析如何结合起来以影响单个车辆的稳定性并改善连接环境中的交通振动，在案例研究中使用了一组在NGSIM数据中经历振动的车辆。在本案例研究中，

我们的分析表明，某些单独车辆的更高稳定性可以减轻排的振荡强度。它还表明，期望的时间间隔和最大加速度是两个有前途的参数，可用于改善单个车辆的稳定性并在连接和/或自动环境中显着平滑排的振荡。特别要注意的是，在考虑我们分析中探讨的所有因素时，调整所需的时间间隔是缓解交通波动最有效的因素。