MODEL AND STUDY OF REAL-TIME DYNAMIC TRAFFIC INFORMATION PREDICTION

At present Intelligent Transportation System -ITS is an advanced topic in the field of transportation. Urban traffic flow guidance system (UTFGS) is the kernel of ITS, which is also called Route Guidance System overseas. UTFGS makes in-vehicle information device show vehicle position, traffic network map, and road situation automatically and it provides the optimal route from the origin to the destination. Since UTFGS can reduce the waiting time of the vehicle and decrease congestion on the street, the traffic volume in the road network will be assigned reasonably. In this case, the operation efficiency of the traffic network will be enhanced. The information of UTFGS can be divided into the static and the dynamic. The static information, such as the scale of the road network, quality, road capacity and the distribution of the infrastructures etc, is obtained and maintained easily. However, it is difficult to check up, process, predict and broadcast the dynamic information, such as traffic volume, travel time, congestion degree, availability of parking etc. Those dynamic information are very significant for the travelers and the traffic administrator. Considering the characteristics of mixed traffic in Chinese cities, in this paper, two real-dynamic traffic information-processing models are presented, including real-dynamic traffic volume combinatorial prediction model and real-dynamic travel time applied prediction model. These two models are introduced.