Xin-Gang Li

Cellular automaton model with time gap dependent randomisation under Kerner's three-phase traffic theory

In this article, a new cellular automaton traffic flow model, in which randomisation effect is enhanced with the decrease of time gap and the long time stopped vehicle has large randomisation probability, is presented. Both periodic and open boundary conditions are considered. Under periodic boundary condition, the fundamental diagram and the properties of synchronised flow are studied in detail. Under open boundary condition, different congested patterns induced by an on-ramp are analysed. The results are consistent with the well-known results of Kerners three-phase traffic theory published before.

CELLULAR AUTOMATA MODEL FOR UNSIGNALIZED T-SHAPED INTERSECTION

In this paper, the unsignalized T-shaped intersection is modeled by a cellular automata model. The main street and the minor street join at the intersection. As to the traffic flow is not controlled by traffic lights, conflict happens between the vehicles from minor street and that from main street. Two different crash avoiding rules are used to dispose the conflicts. In the first rule, the priorities are given to the driving-ahead vehicle and the vehicle on the main street. In the second rule, the vehicle that reaches the conflicting point earlier enters into the intersection. The flux on each lane depending on the inflow rates is studied in detail. The capacity of the system is also investigated. Our simulation results suggest that the two rules do not take the same effect on the capacity under different traffic conditions.

MODELING THE INTERACTION BETWEEN MOTORIZED VEHICLE AND BICYCLE BY USING CELLULAR AUTOMATA MODEL

In this paper, the cellular automata models for motorized vehicle flow and that for bicycle flow are combined to modeling the interactions between the right-turning motorized vehicle and the driving ahead bicycle at intersection. We introduce the probability that the cross point is taken up by the same kind of vehicle during two successive time steps to describe the complex behaviors when conflict happens. The flux of both motorized vehicle and bicycle depending on the inflow rates are investigated and the spatiotemporal diagrams are also presented to show different traffic states as the inflow rates change. The simulation results show that the model can describe the interactions between motorized vehicle and bicycle. It makes foundations for future research on mixed traffic flow.

An effective algorithm to simulate pedestrian flow using the heuristic force-based model

A heuristic method based on the concept of velocity obstacle has been included into the force-based model to describe the pedestrian dynamics. This paper has analysed the continuity and monotonicity of the heuristic function of finding the optimal direction of velocity, and proposed a heuristic detouring algorithm (HDA) to find approximate optimal solutions. In the cases of only one standing pedestrian in the vision field, the solutions calculated by the HDA are optimal; while in other cases, the solutions are near-optimal. To test the performances of HDA, numerical experiments are conducted by three other algorithms that are enumeration algorithm (EA), social force model (SFM), optimal reciprocal collision avoidance (ORCA). Results show that HDA is about 2 times faster than ORCA and 20 times faster than EA. HDA performs with higher efficiency than ORCA and EA, and the fundamental diagram obtained by HDA agrees with empirical data better than ORCA and SFM.

TRAFFIC BEHAVIOR AROUND THE WEAVING SECTION IN CELLULAR AUTOMATA MODEL

In this paper, the traffic behavior around the weaving section, in which an on-ramp and an off-ramp are joined together by an auxiliary lane, is investigated. The drive-in vehicle should merge into the main road while the exit vehicle should diverge from the main road. The influence of those merging and diverging behavior on the dynamics of traffic around the weaving section is studied by analyzing the phase diagram in different spaces and the total flux depending on each parameter. When the number of drive-in vehicle and the number of exit vehicle are close to each other, the total flux is less disturbed. But when they have large difference, the total flux is greatly reduced. We also studied the total flux as a function of the length of the weaving section. It is shown that the weaving section should not be shorter than 150 m.

Characteristics of traffic flow at nonsignalized T-shaped intersection with U-turn movements.

Most nonsignalized T-shaped intersections permit U-turn movements, which make the traffic conditions of intersection complex. In this paper, a new cellular automaton (CA) model is proposed to characterize the traffic flow at the intersection of this type. In present CA model, new rules are designed to avoid the conflicts among different directional vehicles and eliminate the gridlock. Two kinds of performance measures (i.e., flux and average control delay) for intersection are compared. The impacts of U-turn movements are analyzed under different initial conditions. Simulation results demonstrate that (i) the average control delay is more practical than flux in measuring the performance of intersection, (ii) U-turn movements increase the range and degree of high congestion, and (iii) U-turn movements on the different direction of main road have asymmetrical influences on the traffic conditions of intersection.

NETWORK ANALYSIS OF THE EVOLUTION OF TRAFFIC FLOW WITH SPEED INFORMATION

In the cellular automata traffic flow model, the traffic state can be represented by the discrete speed value of vehicles, thus the traffic flow can be deemed as a discrete dynamical system. In the evolution process of traffic flow, complex networks are constructed by representing the traffic state as node and the evolution relationship in timescale as link. The emerging times of link is defined as its weight, then the node strength is equal to the emerging times of the corresponding traffic state. As a result, a weighted network is obtained. The dynamics of stop-and-go traffic are studied by investigating the statistical properties of the network. Simulation results show that scale-free behavior commonly exists in the evolution process of stop-and-go traffic. The degree distribution, node strength distribution and link weight distribution have the power law form. The node with high degree also has large strength. The structure of the network is not influenced by the randomization probability and density as long as the stop-and-go traffic is reproduced.

Empirical Slow-to-Start Behavior from NGSIM Trajectory Data

The slow-to-start rule is usually adopted in cellular automaton traffic flow model. And it is deemed as the mechanism for metastable states and hysteresis effect. In this paper, we explore the slow-to-start behavior by analyzing the vehicle trajectory data provided by the Next Generation Simulation program. Then the slow-to-start rule is verified in several cellular automaton traffic flow models. The results show that the acceleration rate will increases as the speed increasing. The starting up gap does not change while the acceleration rate increases as the stop time becomes larger. The slow-to start behavior really exist, and the slow-to-start rule in VDR model is realistic. But the slow-to-start rule in MCD model is not consistent with empirical results.

Properties of cellular automaton model for on-ramp system

The on-ramp, as a typical bottleneck, has been widely studied by using Cellular Automata model. There are two different kinds of Cellular Automata models for on-ramp. This paper investigate the difference in traffic dynamics between the two kinds of models. The results show that they both have realistic and unrealistic features. The strong points of the two kinds of models should be combined together to model the on-ramp system.

First Order Phase Transition in a New Cellular Automaton Model within the Framework of Kerner’s Three-Phase Traffic Theory

The cellular automata model proposed by Tian and Jia, was established in the framework of Kerner’s three-phase traffic theory and able to reproduce the three-phase traffic flow. But the first order transition from free flow to synchronized flow, which has been observed in the real traffic, could not be reproduced in the Tian and Jia model. One can see that the synchronized flow begin to emerge in Tian and Jia model, when the density is very low, which is obviously not consistent with the real traffic. Therefore, the authors improved their model by the following main aspects. (i) The randomisation probability function and the deceleration extent in the new model have considered more situations according to the real traffic than that of Tian and Jia model. (ii) The slow-to-start rule in the new model is different with that of Tian and Jia model, since the slow-to-start rule in Tian and Jia model is not insistent with the real traffic. (iii) According to the new model, all vehicles will go through the randomisation step first according to its updated rules. Through these improvements, the new model can successfully reproduce the first order transition from free flow to synchronized flow, which is clearly shown on the fundamental diagram. Moreover, one can see the coexistence of the free flow and synchronized flow on the space-time plots, which is strong evidence of this transition. Then the 1-min average flux density diagram, the correlation functions are presented to identify the synchronized flow state. All results have shown that the new model is well consistent with the empirical findings.