车载网是一种以车辆为通信节点的无线自组织网络，旨在实现车与车、车与基础设施之间的数据通信。车辆的高速移动性易引起网络拓扑结构的变化，进而降低数据包的传递率和路由协议的工作效率，甚至导致信道中断。目前，对于车载网通信协议和应用的研究主要借助仿真平台模拟实现，平台内嵌的车辆移动模型性能对协议的分析和研究至关重要。首先，对Simulation of Urban Mobility（SUMO）平台下常用的6种车辆跟驰模型进行了详细的描述；其次，分析并引入影响移动模型性能最明显的3种因素；最终，依托城市道路交通环境，通过设置不同的模拟场景对比分析了在不同跟驰模型作用下的车辆密度、车辆平均速度和道路占用率3个指标。详实的实验结果表明，Krauss模型具有最优异的性能。此外，通过仔细观察单个车辆的跟驰行为从微观上揭示了各模型的工作原理。

Vehicle Ad Hoc Network (VANET) is a kind of wireless ad hoc network, which takes the vehicle as the communication node. It aims to achieve data transmission between vehicle and vehicle, or between vehicle and infrastructure. The high-speed mobility of vehicles can easily cause the change of network topology, reducing the transmission rate of data packets and the efficiency of routing protocol and even leading to channel interruption. The research on the communication protocol and application of the VANET is mainly implemented by the simulation platform. The performance of vehicle mobility model embedded in the platform for protocol analysis and research is crucial. In this paper, firstly, various car-following models under the Simulation of Urban Mobility (SUMO) platform are described in details. Then, three factors that have the most obvious impact on the performance of mobile models are introduced. Finally, based on the urban road traffic environment, the three indicators of vehicle density, average vehicle speed and road occupancy rate under different car-following models are compared and analyzed in different simulation scenarios. The results show that Krauss model has best performance. In addition, simulating the car-following behavior of a single vehicle reveals the working mechanism of each model microscopically.

国家自然科学基金资助项目（61872205）；山东省自然科学基金资助项目（ZR2019MF018）；青岛市应用基础研究计划资助项目（18-2-2-56-jch）。