Research On Short-term Passenger Flow Prediction And Early Warning System Of Urban Rail Transit

The rapid development of urban rail transit has effectively alleviated the sudden increase in demand for public transportation due to population growth.At present,many urban rail transit stations are built underground,with limited internal space and special operating environment,making emergencies,bad weather,morning and evening peaks,etc.,easy to cause congestion problems in the station’s passenger flow.Therefore,how to prevent the possible large passenger flow in advance has become a hot spot of current concern.If accurate passenger flow prediction results are combined with reasonable passenger flow warning judgment conditions,it will have important practical significance for taking preventive measures in advance,alleviating passenger flow congestion,reducing passenger waiting time,and improving ride comfort.This article studies passenger flow forecasting and early warning methods on the basis of fully demonstrating the characteristics and influencing factors of passenger flow.Constructing a short-term passenger flow prediction model based on convolutional long and short-term memory neural network and introducing an optimization model of dynamic inertia weighted particle swarm optimization algorithm,giving the best parameter optimization direction,solving the Conv LSTM model easy to fall into local optimization and gradient dispersion problems,etc.Speed up the training speed of the model.Based on the advantages of the Conv LSTM model,which can extract the temporal and spatial characteristics of passenger flow data,and the internal connection of search data,we have carried out research on the prediction of urban rail short-term passenger flow for different day types and rainfall,and calculated the early warning threshold based on the prediction results to determine the current passenger flow Warning level.The article takes the real historical passenger flow data of the subway as an example,and verifies the efficiency of the model in this article by comparing various models.Through the design of Urban Rail Transit Short-term Passenger Flow Prediction and Early Warning System(USWS),the deep learning prediction model and early warning method are combined in practical application software to upload historical passenger flow Data is used as model input,and passenger flow prediction results,early warning levels and early warning schemes are used as output,in order to help timely management and control of passenger flow in urban rail stations.