一种基于小数据量的快速识别短时交通流混沌特性的方法

A Method of Rapidly Identifying Chaos Characteristics of Short-term Traffic Flow Based on Small-data-sets

对短时交通流进行预测、诱导和控制是智能交通控制系统的重要研究内容。由于对短时交通流进行混沌特性识别时,存在实时性与样本数量之间的矛盾。因此,本文基于混沌时间序列分析理论,提出了一种快速计算短时交通流时间序列最大Lyapunov指数的小数据量方法,用于识别短时交通流中是否存在混沌特性。该方法首先将短时交通流时间序列在相空间中进行重构,以充分提取短时交通流中的相关信息。并结合庞卡来截面法对识别结果进行了验证。从而为对短时交通流进行分析、预测和控制时所采用的相应方法提供了可靠的理论依据。对实测短时交通流行为进行识别的结果表明,该方法具有计算量小、实时性好,对小数据量可靠且容易操作等优点。

The forecasting, inducement and control of the short-term traffic flow is very important for the intelligent transportation control system （ITS）. Because there is a paradox between the real-time and sampling data when we identify the chaotic characteristics of short-term traffic flow, this paper presents a fast and small dataset method for calculating the largest Lyapunov exponent from short-term traffic flow time series based on chaotic time series analysis, which is used to identify whether or not chaotic characteristics exit in the short-term traffic flow. This method reconstructs the time series of the short time traffic flow in the phase space firstly, and the correlative information in the traffic flow is extracted richly. Furthermore, the method＇s results are val idated using the Poincare section. Therefore, this method provides us a reliably theoretical basis for adopting the corresponding method to analyze, predict and control the short-term traffic flow. The identifying results of real short-time traffic flow show that this method has many advantages, such as fast, good real-time, reliable for small data sets and easy to implement.