机非交互路段非机动车越线超车行为建模与仿真

Modeling and Simulation of Cross-Line Overtaking Behavior of Non-Motorized Vehicles at Mixed Flow Road Section

摘要:

为了准确描述机非交互路段的非机动车越线超车行为，在越线超车特征分析及关键指标定义的基础上，提出了基于生存分析的超车时间预测模型，对越线超车行为的持续时间进行预测，并通过实测的160组非机动车越线超车轨迹数据对模型参数进行标定.为验证模型在混合交通流的应用有效性，整合非机动车纵向行驶模型、换道动机模型、间隙选择模型等，将模型应用于团队自主研发的微观交通仿真软件TESS NG中，并与经典的微观交通仿真软件VISSIM进行对比.基于实际路段的仿真分析结果表明，在非机动车越线超车持续时间方面，TESS NG的仿真精度为90.12%，高于VISSIM的67.4%；在中位生存时间方面，TESS NG仿真误差仅为2.56%，优于VISSIM的43.58%.

Abstract:

To depict the cross-line overtaking behavior of non-motorized vehicles at mixed flow road section, this paper analyzed the characteristics of cross-line overtaking behavior and defined the key parameters first, and then proposes the overtaking elapsed time estimation model for non-motorized vehicles based on survival analysis. It utilizes 160 field trajectories of cross-line overtaking non-motorized vehicles to calibrate the proposed model. In order to verify the effectiveness of the model in mixed flow, it further applies the model and integrates it with the car-following model, the lane-changing motivation model, and the gap acceptance model to the self-developed microscopic traffic simulation package TESS NG by adopting the classic micro-simulation model VISSIM as the evaluating benchmark. The result shows that the simulation accuracy of TESS NG is 90.12% in terms of estimating the overtaking elapsed time of non-motorized vehicles, which is far better than the 67.4% achieved by VISSIM and the relative error of median survival time is 2.56% which is 43.58% less than that of VISSIM.