Microscopic Modeling And Properties Analyzing Of Transportation System Based On Cellular Automata Model

Traffic congestion is an inevitable problem to large cities. It makes daily life full of inconvenience, and brings lots of social problems, i.e., traffic accidents, air pollution and energy consuming. Traffic problem restricts the developing of a city to healthy direction. It has been proved by the experience of traffic engineering at home and abroad that traffic planning, management and control should be carried out under the guidance of advanced traffic flow theory, and transportation facilities should be more sufficiently used, then the urban traffic problem can be solved. Modern traffic flow theory was developed as time requires. Since 1990s, cellular automata traffic flow models, with simple rules, good expansibility and high computational efficiency, have risen in traffic flow theory as a new force and gradually win the attention of traffic scholars and engineers. The aim of this dissertation is to modeling and simulation of transportation system with cellular automata model by considering the properties of traffic bottleneck and the characteristics of driver in our country to investigate the evolution dynamics of traffic flow, then give some instructions to traffic engineering. The contents of this dissertation are as follows:1. The empirical vehicle trajectory data are used to analysis the driving behavior, then the consistency between the evolution rules in cellular automata and the empirical driving behavior is verified. It has been shown that:1) the slow-to-start rule in VDR model, CD model and KKW model is consistent with empirical driving behavior; 2) the slow-to-start rule in MCD model is inconsistent with empirical driving behavior; 3) speed adaptation behavior really exists, so the assumption on driving behavior is right.2. A single lane cellular automata traffic flow model considering carefully driving behav-ior and a two lane cellular automata traffic flow model considering aggressive lane changing behavior are proposed.●The carefully driving behavior is modeled by the deceleration in advance rule, i.e., replace the original space gap with the estimated space gap. If there is time delay when estimate the space gap, in intermediate density range synchronized flow can be reproduced with small randomization probability; If there is anticipation when estimate the space gap, the space-temporal diagram of the modified model has little difference with the original model, but the frequency of serious accidents is depressed.●The aggressive lane changing behavior of fast vehicle is introduced into the two lane cellular automata traffic flow model. If the fast vehicle is hindered by the slow vehicle, it will change lane with high probability in the condition that the gap to the back vehicle on the destination lane is larger or equal than 2 cells. In other cases, the vehicle changes lane with low probability. Simulation results show that the "plug" is rarely formed, the flux in the intermediate density range is improved. In the mean time, the lane changing frequency and the ping-pang lane changing frequency are depressed.3. The two common methods for modeling on-ramp system, virtual ramp lane model and acceleration lane model, are compared. The differences of the two methods are given by analyzing the simulation results. At last, the realistic part of the two methods are combined together and a new model is proposed. The evolution dynamics of on-ramp system influenced by the acceleration rate and lane changing behavior are investigated. It is shown that:As the acceleration rate decreases, the flux of on-ramp decreases greatly, however the flux of the main road increases little, then the capacity of the on-ramp system is reduced. As the lane changing behavior becomes more careful, the flux of on-ramp decreases little, however the flux of the main road increases greatly, then the capacity of the on-ramp system is improved. In real traffic, some vehicle have low acceleration rate and the aggressive lane changing behavior really exist, so we believe that the capacity drop of on-ramp system has relationship with low acceleration rate and aggressive lane changing behavior.4. The typical type-A weaving section is modeled by cellular automata model. The evolution dynamics of traffic are analyzed and the influence of length of weaving section, the proportion of weaving vehicle and the lane changing behavior on the capacity are investigated. As the number of weaving vehicle increases, the total flux of weaving section decreases. If the numbers of the two kinds of weaving vehicles are consistent, the disturbance to the main road traffic can be alleviated. In the condition that the traffic on the main road is almost saturated and there are certain number of drive-out vehicles, the total flux will first increase then decrease as the inflow rate of on-ramp increases. In the condition that the traffic on the main road is almost saturated and there are certain number of drive-in vehicles, the total flux will also first increase then decrease as the proportion of drive-out vehicle increases. If the proportion of drive-out vehicle is high, the drive-out vehicle can not drive into the off-ramp in time, then they stop on the main road and have serious influence on the main road traffic. At this time, the capacity of weaving section is greatly reduced. The simulation results also show that the length of weaving section should better longer than 150 m, so as to the main road traffic is less disturbed.5. A cellular automata model for unsignalized T-shaped intersection is established. The influence of left turning vehicles on the dynamics of traffic flow and the capacity under the two crash avoiding rules are analyzed. The main road has two lanes:lane A and B, and the minor road also has two lanes:lane C and D. Lane B is adjacent to the minor road. The left turning vehicles on lane C brings more disturbances to the traffic on lane B than that on lane A, so when the inflow rates of lane A and B are equal, there is no traffic phase in which the traffic on lane A is congested and that on lane B is free. If the proportion of left turning vehicles on lane A and the inflow rate of lane C are fixed, the critical inflow rate and saturated flux of lane A will become larger as the inflow rate of lane B increases when the inflow rate of lane B is small. And if the proportion of left turning vehicle on lane A is small, the crash avoiding rule, in which the priority is given to the vehicle on the main road, makes higher capacity.