面向排放的基于行驶轨迹的交通仿真模型优化

Traffic simulation-based emission estimations have been widely used in traffic environmental evaluation. However, the theoretical defect and estimation error of this methodology have not been evaluated and improved. The key problem is, in the simulated trajectories, the behavior parameters of acceleration/deceleration are not accurately captured, while these parameters are sensitive to vehicle power and emissions. Also, this problem has been proved cannot be solved via existing method of parameter calibration. In this context, multiple instrument sets will be applied to collected massive field trajectories of leading-car’s acceleration/deceleration, car-following and lane-changing behaviors. Numerical simulation will be designed to generate vehicle trajectories, which will be compared with the field ones according to second-by-second speed, acceleration, jerk, and VSP characteristics. The influencing mechanism of simulation models on emission estimation accuracy will be analyzed. The stochastic nature of acceleration will be investigated under physical constrains. A set of simulation models will be optimized based on trajectory analysis, including acceleration model of a leading car, Psycho-Physical (P-P), Crash Avoidance (CA), and Optimal Velocity (OV) car-following models, and mandatory and discretionary lane changing models. In order to guarantee the proposed optimized model improves emission accuracy without sacrificing the performances on the stability for perturbations and the aggregated estimation of traffic parameters, stability analysis for perturbations and consistency analysis with aggregated estimations will be conducted. The proposed models will be applied in case studies and validated through field testing.

通过交通仿真测算机动车排放，已经在交通环境评价中广泛应用。但该方法中的重要理论缺陷和测算误差尚未得到系统的评估与改善。其关键问题是：仿真行驶轨迹中的加减速等功率（或排放）敏感的行为参数未能得到准确刻画。该问题已经被证明难以通过参数标定方法来解决。针对该问题，项目将应用多种设备组合，采集大量实际的车辆加减速、跟驰、换道行驶轨迹和排放数据；建立数值仿真平台，分析模型的轨迹输出与实际轨迹间的逐秒速度、加速度、Jerk、VSP等排放特征指标的差异，剖析仿真模型对排放量化精度的影响机理；剖析物理约束下的加减速的随机分布特性，基于行驶轨迹优化前车加减速模型，生理心理、安全距离、理想速度跟驰模型，以及强制性和判断性换道模型；对优化后的模型进行稳定性分析和集计层参数的一致性分析，在不降低其抗扰动稳定性和集计层仿真性能的前提下，提高微观仿真模型的油耗排放测算精度；通过案例研究，进行模型应用与验证。

通过交通仿真测算机动车排放，已经在交通环境评价中广泛应用。但该方法中的重要理论缺陷和测算误差尚未得到系统的评估与改善。其关键问题是：仿真行驶轨迹中的加减速等功率（或排放）敏感的行为参数未能得到准确刻画。该问题已经被证明难以通过参数标定方法来解决。针对该问题，项目应用多种设备组合，采集了大量实际的车辆加减速、跟驰行驶轨迹和排放数据；建立了数值仿真平台，分析了模型的轨迹输出与实际轨迹间的逐秒速度、加速度和VSP等排放特征指标的差异，剖析了仿真模型对排放量化精度的影响机理；剖析了物理约束下的加减速的随机分布特性，基于行驶轨迹优化了生理心理跟驰模型；对优化后的模型进行稳定性分析和集计层参数的一致性分析，在不降低其抗扰动稳定性和集计层仿真性能的前提下，提高微观仿真模型的油耗排放测算精度；通过案例研究，进行了模型应用与验证。