With the rapid development of vehicle-to-everything communication and autonomous driving technology, research on connected and automated vehicles (CAVs) is experiencing significant growth. Multiple vehicles with different intelligence levels will coexist for the foreseeable future. This paper proposes an adaptive car-following control framework designed to dynamically form platoons or operate individually according to the traffic environment. The aim is to enhance platoon stability, improve efficiency and reduce emissions. Moreover, we consider the stochastic driving behaviors of human-driven vehicles and propose a transposition prediction method that predicts the reaction of rear vehicles to CAV velocity variations from the perspective of rear vehicles. The disturbance scenario and platoon reorganization scenario are designed to conduct comparative experiments with adaptive cruise control, cooperative adaptive cruise control, and distributed model predictive control. The experimental findings underscore the effectiveness of the proposed approach, showing its ability to swiftly and substantially mitigate the impacts of traffic disturbances while simultaneously reducing traffic emissions. Furthermore, the proposed prediction method is identified as a valuable asset for expediting the formation of CAV platoons and enhancing the stability of mixed traffic scenarios.

中文翻译：

---

列队或个人：互联和自动驾驶车辆的自适应跟车控制


随着车联网通信和自动驾驶技术的快速发展，对互联和自动驾驶汽车 （CAV） 的研究正在经历显着增长。在可预见的未来，具有不同智能水平的多种车辆将共存。本文提出了一种自适应跟车控制框架，旨在根据交通环境动态形成队列或单独操作。其目的是增强排队稳定性、提高效率并减少排放。此外，我们考虑了人类驾驶车辆的随机驾驶行为，并提出了一种换位预测方法，从后车的角度预测后车对 CAV 速度变化的反应。干扰情景和编队重组情景旨在进行自适应巡航控制、协作自适应巡航控制和分布式模型预测控制的比较实验。实验结果强调了所提出的方法的有效性，表明其能够迅速、实质性地减轻交通干扰的影响，同时减少交通排放。此外，所提出的预测方法被认为是加快 CAV 列队组建和增强混合交通场景稳定性的宝贵资产。