The modeling of driving behaviors on highways is one of the focal areas in traffic engineering. Currently, there are many highway system models, but most of them are mainly macroscopic .traffic models, which treat the individual cars as the same object and ignore each driver to the whole highway system. However, as we know, the flow on the highway system is the mixed flow with multi-lanes. So, when there are difference in density and velocity of each lane, it becomes easier for cars to change lanes. This action happens even frequently at the on-ramp and off-ramp areas, which may seriously impact the whole traffic flow. So making researches in the microscopic traffic model especially the impact of driver behavior such as car-following and lane-changing to the traffic system is very important. In this paper we propose our highway traffic model with consideration of both micro and macro level models. First we classify, the drivers on the highway as five types, analyze the relationship between the distance of the head and the following cars and the following car's velocity on the highway with three lanes in uncongested flow and congested flow separately; then we use fuzzy reasoning to estimate the percentile of each type of flow; the following we get the macro speed-density relationship, based on it we propose our highway traffic model; in the simulation the data of actual flow, velocity and occupancy are introduced to estimate the parameters in our model, and different simulation results are obtained with different type of mixed flow. At last based on the model and simulation results, we discuss the impact of individual driver behavior to the whole highway system. In the simulation, actual data are used in our model under different traffic situations. We can see that our highway mixed flow model is with both micro and macro level; we can get the impact information of driver behavior to the whole system through changing the percentile of the five driver types; also we can use it to improve the highway performance through forecasting and control based on our model, more according with the real traffic situation then macro-level models. So this model is important for the following investigation in the highway system.
修改评论