平面交叉口的交通控制是交通控制系统理论、方法与实践的重要组成部分，描述平面交叉口实时动态交通需求的模型是实时自适应信号控制的基础依据，本文首先对现有交叉口车流动态特征的研究方法进行综述，在分析车流动态特征产生机理的基础上，指出现有方法的不足，并提出神经网络模型来刻划交叉口动态特征 The traffic control of plane intersections is an important part of the theory, method and practice of traffic control system. The model describing the real-time dynamic traffic demands of plane intersections is the basis of real-time adaptive signal control. Firstly, Based on the analysis of the mechanism of the dynamic characteristics of the traffic flow, this paper points out the shortcomings of the existing methods and proposes a neural network model to characterize the dynamic characteristics of the intersection