交通事故下基于CA-SIR模型的高速公路拥挤传播

CA-SIR Model Based Congestion Propagation of Expressway in Accident Scenario

为探究交通事故对高速公路所造成的拥挤传播影响,将流行病传播SIR模型微观化和动态化。针对其计算求解复杂的问题,将其与元胞自动机(CA)有机结合,并考虑了高速公路事故的特点和影响范围等因素。以单向三车道为例,假设事故占用最左侧车道的4个元胞格,对后续车辆拥堵传播的动态过程进行了模拟。对于模型中关键的拥堵传播率与拥堵恢复率2个参数,则主要通过高速公路上车辆的换道概率和制动概率来设定定量化公式,最终分别得到三者间的确定关系。相关仿真结果表明:所提出的CA-SIR模型与传统的SIR模型相比,在最大拥挤车辆数保持相似的情况下,由于考虑元胞自动机的离散,可以更早反映事故发生后的车道封闭,更符合实际情况;根据相关的拥挤发生时空图,当制动概率增加时,事故带来的拥堵更早发生,更晚消散,拥堵的规模和影响路段长度更大;当换道概率降低时,对车辆最大排队长度及整体行驶时间并无显著影响,但车辆在到达事故区域后所造成的拥挤影响会更早发生,拥挤影响的车辆数也更大。故此,在高速公路交通事故发生后,应通过路侧交通广播、智能网联汽车、车路协同等技术手段,尽可能快地发布交通事故信息,引导车辆提前换道,减少制动概率,以期降低事故影响的拥挤规模和持续时间。

In order to investigate the influence of traffic accident on the congestion propagation on expressway,the epidemic propagation SIR model is combined with cellular automata(CA)for the complexity of its computational solution in a microscopic dynamic way considering the characteristics of expressway accidents and the influence range,etc.Taking one-way three-lane for example,assuming that the accident occupies 4 cells of the leftmost lane,the dynamic process of congestion propagation of subsequent vehicles is simulated.For the 2 key parameters of the model(congestion propagation rate and congestion recovery rate),the quantified formulas are set up mainly through the lane changing probability and braking probability of the vehicles on expressway,and eventually the definite relationship among the 3 factors is obtained.The related simulation result shows that(1)compared with the traditional SIR model,the proposed CA-SIR model can reflect the lane closure after accident earlier and better match the actual situation due to the consideration of CA discretization when the maximum number of congested vehicles is kept similar;(2)according to the related congestion occurrence diagram,when the braking probability increases,the congestion caused by accident occurs earlier and dissipates later,and the scale of the congestion is larger and the length of the affected section is longer;(3)when the probability of lane changing decreases,there is no significant influence on the maximum vehicle queue length and the overall travel time,but the influence of congestion caused by vehicles arriving at the accident area occurs earlier and the number of vehicles affected by congestion is larger.Therefore,after the occurrence of traffic accident on expressway,traffic accident information should be released as quickly as possible through roadside traffic broadcasting,intelligent networked vehicles,vehicle-infrastructure cooperation and other technical means to guide vehicles to change lane ahead of time and to reduce the braking probability,so as to reduce the scale and duration of congestion affected by the accident.