Passenger Behavior And Cooperative Guidance Evacuation Under Typical Scenarios Of Stations

Rail transit has become the aorta of the national economy and public transportation in China,and its operating mileage and passenger volume are the highest in the world.As the operating environment becomes more complex,rail transit presents new features such as the normalization of ultra-large passenger flow and the complexity of the mechanism of delayed propagation.Facing the operation demand of huge passenger flow,once the passenger flow exceeds the carrying capacity of the station,it is easy to cause passenger congestion and even lead to safety accidents such as crowd trampling.Typical scenarios such as station platforms,stairs,and escalators are areas where passengers are highly dense.The intersection of passenger movements,conflicts of opposites,and unbalanced path selection can easily cause congestion.The traffic efficiency of passengers and safety evacuation in stations have attracted wide attention from scholars at home and abroad.Aiming at these typical scenarios,the internal laws of passengers' macroscopic movement under different scenarios and the relationship between the microscopic characteristics of passengers' movements and macroscopic phenomena are explored.It is the research focus of improving passenger traffic efficiency and ensuring passengers 'safe evacuation.And it has important theoretical and practical significance to reduce or even avoid accidental casualties.Therefore,this paper focuses on the dynamical characteristics and behavior prediction of passengers under typical scenarios of stations and the emergency evacuation of passengers based on cooperative guidance.The main research is summarized as follows:Firstly,the evolution mechanism of the passenger motion is given,according to the dynamical characteristics of passengers on the stair,and a hybrid model of passenger motion is constructed.Analyze the structural particularity of the scenario and the movement characteristics of passengers,and a hybrid model that simulates the movement of passengers on the stairs is established.The validity of the model is proved based on the data fitting of the passenger's fundamental diagram and the recurrence of the movement phenomenon.On this basis,the dynamical characteristics of passengers such as the average speed and passenger flow are further studied.And the evolution mechanism of the proportion of slow crowds in unidirectional passenger flow and the proportion of opposing crowds in bi-directional passenger flow on passenger dynamical characteristics are also described.Secondly,aiming at the problem of low capacity of stair and escalator mixed area caused by imbalanced passenger routing,the prediction method of passenger path selection behavior based on a neural network model is proposed.A neural network model for the prediction of path selection behavior is established.And the nonlinear relationship between different factors and passengers' path selection behavior is characterized.The passenger path selection is predicted based on the neural network model,and the average influence values of the factors are given to provide a theoretical reference for the optimization of passenger flow control strategies.A computational experiment simulation module is established,and the effectiveness of the proposed passenger flow control strategy to optimize the capacity of the bottleneck area is verified.The research results can provide a theoretical basis for the optimization of passenger organization and improve the traffic efficiency of the bottleneck area.Finally,for the evacuation of platform passengers in emergencies,the emergency evacuation strategy for platform passengers under cooperative guidance is proposed,and the evacuation paths of leaders are optimized based on reinforcement learning.An optimized model of passenger emergency evacuation strategies based on evacuation cost,evacuation efficiency,and safety is established.The number and location allocation strategies of different guidance methods are given based on the model.Based on reinforcement learning,the process of pathfinding for the leader is established,and the optimization of passenger evacuation path under the guidance of the leader is given.A passenger evacuation experiment simulation module is established,and the effectiveness of the proposed collaborative guidance strategy and evacuation path to improve the efficiency of passenger evacuation are verified.The research results can provide corresponding guidance for the formulation of emergency evacuation plans for stations in emergencies.