Recognition and Analysis of Traffic Congestion Level

Conventionally, the congestion state of a freeway is judged by mean velocity, occupancy, flow or travel time, but most of them are difficult to apply under variable conditions. And AC algorithms such as a California algorithm, McMaster algorithm etc., can detect the congestion accidents, but cannot calculate the congestion level successively. A new model considering the the mean traffic velocity, the vehicle proportion of mixed traffic flow, the geometry of the road, and the weather conditions is proposed to describe the traffic congestion level uninterruptedly. Moreover, the traffic congestion states are defined based on the new model. The results of the experiment indicate that the new model is ideally suited for inspecting the traffic state in real-time.