【摘要】 文中基于低频和稀疏的GPS轨迹数据提出一种新的信号控制参数估计方法.基于车辆轨迹数据周期性聚合中存在的分布模式差异，提出一种投影累计密度曲线法估计信号周期，并基于蒙特卡罗模拟得到正确估计周期所需的最小样本量.在时距图中利用交叉口不同相位的消散冲击波与对应流向进口道停车线的交点的相对位置关系，估计信号交叉口相位组织，以及各相位的有效绿灯时间.通过实际和仿真数据综合进行验证，结果表明：所建立的模型仅用几十条低频轨迹数据便可获得较高的估计精度，优于现有方法.更多还原

【Abstract】 A new estimation method of signal control parameters based on low frequency and sparse GPS trajectory data was proposed. Based on the difference of distribution patterns in periodic aggregation of vehicle trajectory data, a projection cumulative density curve method was proposed to estimate the signal period, and the minimum sample size required for correct estimation of the period was obtained based on Monte Carlo simulation. In time-distance graph, the phase organization of signalized intersection and the effective green time of each phase are estimated by using the relative position relationship between the intersection of the dissipated shock wave in different phases and the corresponding stop line flowing to the entrance. Through the comprehensive verification of actual and simulation data, the results show that the established model can obtain higher estimation accuracy with only dozens of low-frequency trajectory data, which is better than the existing methods.更多还原