A Study on Capacity of Weaving Sections

本文通过对都市高速公路的若干观测，探讨了高速公路路段的编织能力。本研究分为两部分，一是宏观分析织造路段的交通流量、速度、密度等交通流状况，二是微观分析每辆车的运动，以建立能够再现织造能力的仿真模型。宏观分析表明，部分编织路段的通行能力比简单路段的通行能力低26%，且其通行能力随公路几何形状和编织长度、体积比等交通条件的不同而变化。其次，将1965-HCM（公路通行能力手册）估算的编织影响因子值与观测值进行比较。发现估计值略小于观测值。第三，对现有的四种织造速度预测方法的适用性进行了检验；这些是1985-HCM， JHK， PINY， a…More和Leisch方法。1985-HCM方法与观测到的速度最吻合，而PINY和Leisch方法往往低估了速度。在微观分析方面，利用观测得到的车辆轨迹数据，分析了编织车辆的相对位置和相对速度等微观变量。在此基础上，建立了估算织造能力的仿真模型。认为编织车辆总是选择与目标间隙相对应的最佳加速度和所需变道时间，并对这些变量进行调整，直到编织完成。在这个模型中，每辆车通过预测自己在编织完成时的状况，随时精确地更新这种特定的动作。仿真模型在两个拥堵的都市高速公路交叉路段进行了验证。结果表明，容量和速度的模拟值与实测值吻合较好。最后讨论了不同交通条件和几何条件下的编织路段通行能力。少

This study discusses capacity of weaving sections based on several observations on the Metropolitan Expressway. The study is divided into two parts : a macroscopic analysis of traffic flow conditions such as flow, speed, and density of traffic in a weaving section, and a microscopic analysis of each vehicle motion in order to develop a simulation model which can reproduce weaving capacity.For the macroscopic analysis, we first show that capacities of some weaving sections are lower than those of simple section by at most 26% and that the capacities vary depending on highway geometry and traffic conditions such as weaving lengths and volume ratios. Secondly, we compare values of weaving influence factor estimated by the 1965-HCM (Highway Capacity Manual) with the observed values. The estimates were found to be slightly smaller than the observed values. Thirdly, we examine applicability of four different existing methods for predicting weaving speed ; those are the 1985-HCM, JHK, PINY, a … More nd Leisch methods. The 1985-HCM method gives the best agreement with the observed speed but the PINY and Leisch methods tend to underestimate the speed.For the microscopic analysis, using vehicle trajectory data obtained from observations, microscopic variables such as relative position and relative speed of weaving vehicles are analyzed. Based on these data, a simulation model estimating weaving capacity was developed. It is considered that weaving vehicles are always selecting an optimum acceleration and the required time for lane changing corresponding to the aimed gap, and they adjust these variables until the weaving is completed. In this model, every vehicle momentarily updates such particular maneuvers precisely by predicting its condition at the moment the weaving is completed. The simulation model was validated at two congested weaving sections on the Metropolitan Expressway. As a result, the simulated values f capacity and speed fit well with the observed ones. Finally capacities at the weaving sections under various traffic and geometric conditions were discussed. Less