Pedestrian Flow Prediction in Open Public Places Using Graph Convolutional Network

Open public places, such as pedestrian streets, parks, and squares, are vulnerable when the pedestrians thronged into sidewalks. The crowd count changes dynamically over time with various external factors, surroundings, weekends, peak hours, so it is essential to predict accurate timely count. To address this issue, study introduces graph convolutional network (GCN), a network-based model, flow in walking street. Compared other grid-based methods, model capable of directly processing road graphs. Experiments show GCN its extension STGCN consistently significantly outperform five baseline models, namely HA, ARIMA, SVM, CNN LSTM, terms RMSE, MAE R2. Considering computation efficiency, standard was selected crowd. results showed that obtains superior performances higher prediction precision on weekends which R2 above 0.9, indicating can capture features effectively, especially during periods massive crowds. will provide practical references for city managers alleviate congestion help make smarter planning save travel time.