Abstract: For the complex mixed traffic flow composed of four types of vehicles: human-driven vehicle (HDV), connected human-driven vehicle (CHV), autonomous vehicle (AV) and connected autonomous vehicle (CAV). The impacts of driver’s reaction delay, vehicle communication delay and sensor measurement delay on the stability for complex mixed traffic flow are considered all at once. Fundamental diagram and stability models are constructed considering multiclass time delays for complex mixed traffic flow. Firstly, values of the proportion of different car following types and the concerning time delay for complex mixed traffic flow are analyzed, while the fundamental diagram models of homogeneous and complex mixed traffic flow are derived and analyzed. Then, the stability condition considering multiclass time delays for complex mixed traffic flow is derived theoretically, and stability for complex mixed traffic flow under different penetration rate of CHV and CAV is analyzed in detail. At last, Matlab numerical simulation experiment is designed to analyze the impacts of driver’s reaction delay, vehicle communication delay and sensor measurement delay on stability for complex mixed traffic flow in detail. The results show that: (1) The increase in penetration rate of CAV can effectively improve the capacity and stability of mixed traffic flow, with a significant increase compared to the increase in penetration rate of CHV. (2) When the penetration rate of CAV is higher than 0.6, the complex mixed traffic flow remains stable. (3) All types of time delays have a negative impact on the stability for complex mixed traffic flow. The driver’s reaction delay has the greatest impact on the stability for complex mixed traffic flow, while the sensor measurement delay has the smallest impact on the its stability, and the impact of vehicle communication delay on its stability is in between the two.

Key words: intelligent transportation, fundamental diagram and stability, car following model, complex mixed traffic flow, multiclass time-delay

摘要： 针对由人工驾驶车辆（human-driven vehicle，HDV）、网联人工驾驶车辆（connected human-driven vehicle，CHV）、自动驾驶车辆（autonomous vehicle，AV）和网联自动驾驶车辆（connected autonomous vehicle，CAV）四种类型车辆构成的复杂混合交通流，同时考虑驾驶员反应时延、车辆通讯时延以及传感器量测时延对复杂混合交通流稳定性的影响，构建了考虑多类时延的复杂混合交通流基本图与稳定性模型。对复杂混合交通流中不同跟驰类型的比例与其时延的取值进行了分析，同时对同质流与复杂混合交通流基本图模型分别进行了推导与解析。理论推导出了考虑多类时延的复杂混合交通流稳定性条件，并详细解析了不同CHV与CAV 渗透率下复杂混合交通流的稳定性。设计了Matlab数值仿真实验，详细分析了驾驶员反应时延、车辆通讯时延以及传感器量测时延对复杂混合交通流稳定性的影响。结果表明：（1）智能网联车辆渗透率的提高能够有效提高复杂混合交通流的通行能力与稳定性，同比网联人工驾驶车辆提升幅度要大；（2）在CAV渗透率高于0.6的情况下，复杂混合交通流迅速达到稳定状态；（3）各类时延对复杂混合交通流的稳定性均具有消极影响，其中驾驶员反应时延对复杂混合交通流的稳定性影响最大，而传感器量测时延对其稳定性的影响最小，车辆通讯时延对其稳定性的影响处于二者之间。

关键词: 智能交通, 基本图与稳定性, 跟驰模型, 复杂混合交通流, 多类时延