Short-Term Urban Rail Passenger Flow Prediction using TCN-LSTM Model

Abstract

Predicting short-term passenger flow in urban rail transit systems is crucial for efficient operation and management. The TCN-LSTM model, which combines Temporal Convolutional Networks (TCNs) and Long Short-Term Memory (LSTM) networks, offers a robust solution for this task. TCNs provide an expansive receptive field, enabling the model to capture long-range dependencies and trends in passenger flow data. Meanwhile, LSTMs are adept at handling sequential data, refining the temporal features extracted by the TCN layers to account for short-term dependencies and fluctuations. This hybrid approach leverages the strengths of both architectures, resulting in a model that can accurately predict passenger volumes, accommodating regular traffic patterns and sudden changes in rider behavior. The implementation of the TCN-LSTM model in urban rail transit systems leads to optimized train schedules, improved service efficiency, and enhanced passenger experiences. By enabling more dynamic and responsive transit management, this model plays a pivotal role in addressing the challenges of modern urban transportation.