Research On Characteristics Analysis And Control Method Of Urban Rail Transit Passenger Flow

With the continuous expansion of urban rail transit network and the continuous growth of passenger travel demand,the problem of passenger congestion is becoming increasingly serious,which reduces the traffic efficiency and operation stability of the network.When there is a large passenger flow,it mainly depends on the subjective experience of the managers,lacks the accurate grasp and analysis of the passenger group aggregation characteristics and the distribution state of the line network passenger flow,and lacks the effective guidance and control measures for the passenger flow congestion.Therefore,solving the above problems is conducive to improving the organization and management ability of passenger flow under the condition of network operation,and providing auxiliary decision-making for emergency disposal of urban rail transit.Based on this,combined with the current situation of urban rail transit passenger flow operation and management under the network operation condition,this paper takes "passenger flow characteristic analysis-passenger flow state discrimination-passenger flow distribution prediction-passenger flow control" as the technical route of urban rail transit.The specific research contents include:(1)Urban rail transit station passenger group aggregation mode mining: To explore and master the spatial-temporal variation law of passenger group aggregation in urban rail transit station plays a positive role in optimizing vehicle scheduling between urban rail transit network,especially in optimizing passenger flow organization and management under disaster conditions.Based on WIFI positioning data,this paper studies the aggregation and distribution characteristics of passenger flow in urban rail transit stations.According to the uneven distribution characteristics of passenger positioning data,a data stratification method based on Gaussian mixture model is proposed,and the DBSCAN algorithm is improved to solve the problem that it can not be directly applied to the passenger positioning data with uneven distribution characteristics in urban rail transit stations.Through the calculation of standard and measured data,the advantages of the improved algorithm are proved in terms of clustering accuracy and clustering effect(anti noise ability and real-time performance).It shows that DBSCAN algorithm optimized based on Gaussian mixture model has better clustering effect for non-uniform passenger location distribution data.By understanding the station aggregation characteristics of passengers,it provides a feasible basis for subsequent passenger flow state early warning and control management.(2)Identification and control of passenger flow status in urban rail transit stations: Aiming at the problem of large passenger flow congestion in urban rail transit stations during peak hours,the dynamic control of passenger flow in stations is carried out.Firstly,the cloud model is constructed to judge the status of passenger flow in the station.When the control threshold is reached,the inbound flow is controlled.Based on the linear quadratic optimal control theory,the passenger flow feedback control strategy is designed.By establishing the discrete state space equation of station passenger flow and combining with the station service level,the optimal state feedback controller is designed to obtain the optimal passenger flow control sequence,and the influence of weighting matrix on the control sequence is analyzed,so as to minimize the overall difference between the actual passenger flow density and the expected passenger flow density in the control period.Taking a typical transfer station as an example,the simulation results show that the linear quadratic algorithm can make the control density recover quickly and stabilize at the expected value,the fluctuation of passenger flow is small in the control process,and it can deal with the disturbance of general transfer passenger flow.(3)Prediction and simulation of passenger flow distribution in urban rail transit network: firstly,short-term od passenger flow prediction is carried out to determine the input characteristics of the prediction model and the optimal time interval for short-term od passenger flow prediction,and a multi input and multi output prediction framework for the whole network is constructed;secondly,a passenger route matching algorithm is constructed to transform the passenger route matching problem into a classification problem,and a method based on Bayesian network is proposed Finally,a passenger flow distribution prediction system is constructed to analyze the simulation operation logic of passengers and the main body of the train in the simulation system,and a simulation scenario is built to verify the results.(4)Collaborative control of passenger flow in urban rail transit network: for line level passenger flow control,a passenger flow collaborative control method based on containment control theory is proposed.The method aims at minimizing the difference between the actual density and the expected density of each station in the line,designs a multi station collaborative control strategy,and judges whether the passenger flow is triggered by comparing the passenger flow characteristics and the carrying capacity of each station.Based on the analysis of the controllability of the passenger flow status of urban rail transit lines,the stability conditions of the passenger flow status are derived to select the key controlled stations and their feedback gain matrix.The feasibility of the method is proved by case analysis.