Dynamic Modeling And Phase Transition Of Pedestrian Flows With Consideration Of Subconscious Effects And Behavioral Analysis

Recently, various accidents caused by pedestrian jams have been increasingly happening, which have led to enormous economic losses and posed a serious threat to human life. It is an urgent issue to manage scientifically and control effectively pedestrian jams. Because the pedestrian system is a self-driven many-particle complex system, pedestrians not only have strong nonlinear interactions among themselves, but also have many significant characteristics, such as independence, randomness, flexibility and compressibility. Compared with the study of vehicle flows, coping with the pedestrian flow is more complicated and challenging. Therefore, it is of great importance to establish reasonable pedestrian flow models to investigate macroscopic characteristics and jamming mechanism of the pedestrian flows, which will be helpful for improving the basic theory and application level of pedestrian traffic. In this dissertation, based on the existing macroscopic and microscopic models for pedestrian flows, several improved mathematical models are proposed to investigate pedestrian counter flows, single-file dense crowd flows and multi-type pedestrian queue by considering microscopic characteristics and behavioral differences of pedestrians, respectively. And the macroscopic dynamic characteristics of pedestrian flow and the formation mechanism of pedestrian jams are studied via theoretical analysis and numerical simulation. Moreover, we put forward some suggestions for the planning of pedestrian facilities, and management and control of pedestrian flows.The main contents of the dissertation are as follows.â… . From a viewpoint of behavioral modeling, two extended lattice gas models are established to simulate pedestrian counter flows by considering the human subconscious behavior, the different maximum velocities and the variable transition probability, and the macroscopic dynamic characteristics of pedestrian flow and the formation mechanism of pedestrian jams are investigated.Based on the original lattice gas model proposed by Muramatsu et al, a modified model is proposed to simulate pedestrian counter flow by considering the human subconscious behavior and different maximum velocities. The pedestrian dynamics in the different scenarios, such as various strengths of the preferential direction of subconscious behavior, different combinations of maximum velocities, symmetric and asymmetric injection rates, are investigated under the periodic and open boundary conditions, respectively. The simulation results show that the subconscious behavior can strengthen the stabilization of system and reduce the occurrence of jams effectively. And at larger mixed velocity, the obtained critical transition point corresponding to the maximum flow rate is in good agreement with empirical results. Furthermore, the improvement of movement efficiency mainly depends on the magnitude and asymmetry of entrance density. In addition, to better quantitatively describe the different subconscious behaviors and distinguish their strength variations, an improved lattice gas model with variable transition probability is established by introducing a strength function, which reflects the pedestrian preferential walking habits. And the influence of pedestrians of violating the traffic regulations and the jamming mechanisms are discussed in detail. It is found that the jamming cluster is mainly caused by the pedestrians of violating the traffic regulations. In other words, the pedestrians of obeying the traffic rules could effectively decrease the occurrence of jam cluster. In fact, the two extended models can reproduce some essential features of pedestrian counter flow, such as the lane formation, segregation effect and phase separation. All these indicate that the presented modeling ideas and methods can be contributed to the development of the pedestrian flow theory.â…¡. An improved lattice gas model is proposed to simulate pedestrian counter flow by taking into account the different strengths of following behavior, and the underlying mechanisms of macroscopic phenomena caused by the following effect are studied.A modified lattice gas model is presented to investigate the pedestrian counter flow by considering the different following behaviors under the different walking environments. The influences of system size, asymmetrical conditions, traffic regulation violating behaviors and the strengths of subconscious behaviors on pedestrian counter flow are investigated via numerical simulation. Furthermore, the characteristics of space-time dynamics, phase transition and "deadlock" problem are also discussed in detail, respectively. The simulation results show that the presented model can capture some essential features of pedestrian counter flows, e.g., the lane formation and segregation effect. When the width and the length are larger, the jamming density is independent of system size, and the larger asymmetrical conditions can not only strengthen the stability of system and reduce the occurrence of the jam, but also better solve deadlock problem. In addition, it is also found that the mistaken navigation of the pedestrians of violating the traffic regulations and the blind following of other pedestrians are the underlying mechanism for inducing the jams. Finally, some practical suggestions are given for optimizing pedestrian facilities and preventing pedestrian jams.â…¢. Based on the pedestrian flow experiments, an extended optimal velocity model is established to simulate unidirectional dense crowd by considering asymmetric interaction force, and the jamming mechanism and the nonlinear properties of the density waves are investigated.Based on the experimental observation, an extended optimal velocity model is proposed to simulate single-file pedestrian movement at high density by introducing an asymmetric interaction function, which reflects the differences of interacting forces (i.e., attractive force and repulsive force) between pedestrians. The stability criterion and the neutral stability curves are obtained by using the linear stability theory. The comparison and analysis of the phase diagrams show that asymmetric effect plays an important role in stabilizing the system. The modified Korteweg-de Vries (mKdV) equation is derived through applying the nonlinear analysis. The pedestrian jam could be described by the kink-antikink soliton solution for the mKdV equation. Through the simulation of space-time evolution of the pedestrians' headway, it can be found that the model can suppress effectively small perturbations. Furthermore, the simulation results are consistent with the theoretical analysis and can better reproduce experimental phenomena (i.e., delay the occurrence of stop-and-go). All these indicate that the proposed model provides a new approach to elucidate the formation and evolution mechanism of pedestrian jams in dense crowd, and preventing and controlling the occurrence of pedestrian jams.â…£. A heterogeneous optimal velocity model is constructed to simulate multi-types pedestrian queue movement, and the evolution rules of pedestrian movement and the formation mechanism of space-time pattern are studied.Based on the optimal velocity model proposed by Nakayama et al, a heterogeneous optimal velocity model is established to simulate multi-types pedestrian queue movement by considering the differences among pedestrians (e.g., the differences in age and sex, etc.) and the asymmetric interaction. Through using the linear stability theories, the stability criteria under the periodic and open boundary conditions are obtained. For three kinds of different configurations on pedestrian queue, the evolution rules of pedestrian movement and the formation mechanism of space-time pattern under two different boundary conditions are investigated. The numerical simulations show that the model can better depict the space-time distribution of pedestrian movement from microcosmic viewpoint and reproduce such important phenomena as collective behaviors, dispersion of jams and cluster movements.â…¤. Based on two-dimensional bidirectional pedestrian lattice hydrodynamic model, an extended model is proposed to simulate two-dimensional pedestrian flows by considering the scope of visual field, and the influence of visual field effect on the system stability are investigated from the macroscopic viewpoint.Through incorporating the concept of one-dimensional cooperative driving intelligent transportation system and considering the influence of the behavior of more pedestrians ahead (i.e., the scope of visual field), a modified two-dimensional lattice hydrodynamic model is established to investigate theoretically bidirectional pedestrian flow. The stability criterion and the neutral stability curves are obtained by using the linear stability analysis and the mKdV equation describing the density waves in pedestrian congestions are derived through nonlinear analysis. The results show that utilizing the information of more pedestrian in front can lead to the stabilization of pedestrian systems, and reduce the occurrence of pedestrian jams. Furthermore, through discussing the qualitative relationship between the critical sensitivity and other parameters, a method to enhance the stability can be acquired, i.e., jams can be suppressed by controlling the densities in different directions reasonably.The final chapter of this dissertation is devoted to a summary of the thesis and a prospect of further study of the pedestrian flows.