论文部分内容阅读
受数学模型发展的限制,以往关于交通流状态辨识方法的研究主要侧重于预辨识(交通流预测)及实时辨识(事件检测或交通流质变检测),忽略了对交通流演化规律的研究。为有效改善道路交通安全环境,根据交通流的分布特点,结合交通流在自由流状态、拥挤流状态及间歇流状态的分布均可由Γ分布表达的理论,采用英国南安普敦大学交通研究团队提供的交通流数据,运用最小二乘法及分布变点检验算法对观测数据进行变点搜索及检验。实例验证表明:所建立的基于Γ分布检验变点的交通流演化辨识方法,无需提前确定交通流所服从的具体分布,就能检测出交通流变点是否存在。
Due to the limited development of mathematical models, previous researches on traffic flow state identification mainly focus on pre-identification (traffic flow prediction) and real-time identification (event detection or detection of traffic flow qualitative change), neglecting the study of traffic flow evolution. In order to effectively improve the traffic safety environment of road, according to the distribution characteristics of traffic flow, the distribution of traffic flow in free flow state, crowded flow state and intermittent flow state can be expressed by Γ distribution theory and by the traffic research team provided by Southampton University Traffic flow data, the use of least square method and distributed variable point inspection algorithm to change the point of observation data search and test. The example verification shows that the established traffic flow evolution identification method based on the Γ distribution test can detect the existence of the traffic rheological point without determining the specific distribution obeyed by the traffic flow in advance.