关键词: 城市交通, 交通流, 预测, 混沌理论, 耗散系统, 多元局域预测法

Abstract: With the ever-increasing motor vehicle population, traffic congestion is a severe problem of urban traffic. Traffic flow forecasting may help the traffic management branch to formulate relevant policies, optimize traffic management and solve the traffic problem, at last release the traffic pressure in a certain extent. It is one of the important issues of intelligent transportation systems. Short-time traffic flow forecasting is the main technology of traffic control and vehicle-based navigation. This paper points out that transportation system is a dissipative system and the chaos exists in traffic status. It restores dynamical systems with chaos and fractal theory and predicts the traffic flow using multivariate time series of local prediction method. The basic traffic data are collected to test the effectiveness of the model, and the evaluation index of different time interval of the traffic flow series is analyzed. The proposed method has high prediction accuracy within 2 or 5 minutes level.

Key words: urban traffic, traffic flow, forecasting, chaos theory, dissipative system, multivariate local method