区域交通流特征提取与交通状态评估方法研究

随着城市各种交通控制系统(UTC)的广泛应用,交通控制中心能够采集大量的交通流数据。通过对这些数据的分析来评价区域的交通运行质量和路口状况,从而有助于改进交通控制方案,这是一项新的研究内容。本文通过对区域内路口大量的检测器数据的分析,利用K均值聚类和EM算法相结合的方法,提取出表征路口交通流运行状况的特征向量。然后,给出典型交通区域内路口交通流运行的平衡因子,建立评价区域内路口相似性的相似性矩阵。利用北京市两个典型交通区域的实际交通流数据进行验证,结果表明该评估模型能够很好的反映北京市交通状况,并具有较好的可移植性。

Study on the Method for Regional Traffic Flow Feature Extraction and Traffic Status Evaluation

Along with the wide application of the Urban Traffic Control (UTC) system, traffic control center can collect a mass of traffic flow data. Analyzing on these data can be used to evaluate the region's traffic operating quality and intersection's traffic operating status,to improve the traffic control strategy and bring a new research field. With K-means clustering method and Expectation Maximization (EM) Algorithm,an eigenvector was extracted to describe the traffic status of an intersection by analysis on a mass of data from loop detectors. According to the analysis of two typical regions,equilibrium factors were obtained for evaluating the intersections' traffic flow and the similarity matrix for evaluating the degree of similarity between intersections. Case study in two typical regions of Beijing proved that the proposed models can reflect the real traffic status perfectly and have good transplantability.