Abstract: Link average speed is one of the most important traffic parameters for urban traffic management. In this paper a new link average speed estimation model based on probe data is discussed. After anglicizing the sampling size of the probe for link average speed estimation, the exponentialy smooth model with adaptive weight is built. Because of the link average speed estimation easily influenced by the number of probe vehicles, error tracking signal and probe vehicle number index are introduced to the weight. Then VISSIM simulation on main road is used to evaluation the model and the result shows that even the probe penetration cannot satisfy the requirement of the sampling size, comparing to the arithmetic mean method, the new exponentialy smooth model of this paper can effectively decrease the error and variance of the link average speed estimation.

Key words: probe vehicle, link average speed estimation, adaptive weight exponentialy smooth, VISSIM simulation

摘要： 在讨论探测车样本数量统计需求的基拙上，提出了一种基于探测车技术的路段平均速度自适应加权指数平滑佑计模型，模型中加权系数随着跟踪误差大小和路段上探测车数量的变化而自适应调整，并利用VISSIM仿真数据验证了该模型。结果表明：在路段上探测车占交通流比例随时间变化且不能充分满足统计要求的情况下，相对传统的算术平均方法，基于指数平滑的路段平均速度佑计能够有效提高估计精度，并且可有效减小由路段上探测车数量变化而引起的速度佑计方差。

关键词: 探测车技术, 路段平均速度估计, 自适应权重指数平滑, VISSIM仿真