���п���·�ϻ������ȹ��ʷֲ�������ģ��

在交通油耗和排放模型的研究中，利用机动车比功率（VSP）分布刻画交通状态成为了最新研究需求. 然而，现有交通工程还缺乏对常见交通参数与VSP间关系的认识，更缺乏利用交通参数求解VSP分布的模型. 为了研究交通状态对VSP分布的影响，本文利用大量城市快速路上的浮动车数据，建立了其不同行程速度下对应的VSP分布. 经分析，发现了VSP分布与平均行程速度之间的规律性特征：当平均行程速度大于20 km／h时，VSP分布近似于正态分布；分布均值为以该速度匀速行驶时的VSP值；分布标准差可表达为平均行程速度的幂函数. 基于以上发现，提出了利用平均行程速度的VSP分布数学模型，并利用该模型进行机动车油耗测算. 通过与实测油耗的对比分析，VSP分布模型可以有效用于机动车油耗测算. 本研究指出了利用数学模型描述不同交通状态下VSP分布的可能性，该模型可有效地与交通数据或模型结合，实现油耗和排放的实时量化评价.

Distribution Characteristics and Models of Vehicle Specific Power on Urban Expressways

The development of new fuel consumption and emission models creates the need to characterize traffic conditions by using the Vehicle Specific Power (VSP) distribution. However, in the existing transportation engineering, there has been the lack of knowledge of relationships between the VSP distribution and the commonly used traffic parameters, and the lack of models to develop the VSP distribution from traffic parameters. In order to examine how traffic conditions impact the VSP distributions, this study uses large samples of floating car data collected from the expressways in Beijing to associate the VSP distributions with various average travel speeds. After a comprehensive analysis, regular patterns are found between the VSP distribution and the average travel speed. Specifically, when the average travel speed is higher than 20 km/h, the VSP distribution comes close to a normal distribution. The mean of the VSP distribution is the VSP value when cruising at the average travel speed, and the standard deviation could be expressed as a power function of the average travel speed. Based on these findings, a mathematical model for developing the VSP distributions is then derived by using the average travel speed. Finally, a comparative analysis between the estimated and actual fuel consumption demonstrates that the VSP distributions developed by the proposed model are applicable for fuel consumptions estimations. This study indicates a possibility of developing VSP distributions mathematically for different traffic conditions, which can be integrated efficiently with traffic models or data for the real\|time estimation of fuel consumption and emissions.