处理交通流不确定性问题的理论与方法研究

In this project, the new viewpoints, the new conception, and the new classification for the uncertainty of traffic flow are presented. It is pointed out that the essential uncertainty of traffic flow is deferent from the randomness of traffic flow and the essential uncertainty exists in traffic flow generally. The following approaches to deal with uncertainty are proposed: the approach for the short-term (shorter than 5 min) predictability of traffic flow based on fractal theory; the approach for the short-term forecasting of traffic flow based on wavelets analysis; the approach for detecting of bicycle flow based on pattern recognition. The dynamic response under deferent conditions for an intersection is studied. The AI-based dynamic route guidance strategy is proposed. An approach using Erlang random distribution which parameters change also randomly is proposed to describe the essential uncertainty of traffic flow. The simulation models for generation and movement of traffic flow are constructed based on the approach. The microscopic simulation system for traffic flow with uncertainty is developed. The system is a powerful tool for study on traffic flow, for transportation planning and for traffic control because it simulates the traffic flow with uncertainty more correctly than other systems.

交通流过程是典型的不确定性过程，无论研究交通流理论本身，还是研究交通控制与交通诱导，交通流的不确定性都是不可回避的实质问题与难题。本申请研究交通流的不确定性及其处理的理论方法，为提高交通控制与诱导的有效性提供理论基础，开发出的仿真软件将是研究有效的交通控制与交通诱导策略的基本工具，从而促进ITS的发展。