Large-Scale Indoor Pedestrian Density Prediction Based on Neural Network Model

A new indoor crowd density prediction algorithm suitable for large-scale indoor pedestrian flow was proposed. Based on the current crowd density algorithm with wireless signal intensity, a weighted operation was introduced to improve the estimation quality. Further, according to the estimated human flow density in several consecutive time periods, the BP neural network model is used to predict the crowd density in this area at a certain time in the future.According to the simulation model and the data collection and analysis of three months, the accuracy of the prediction model can reach 94.70%.