Simulation of Evacuating Crowd Based on Deep Learning and Social Force Model

Xin Li,Yu Jiang,Minghao Zhao,Hongtao Bai,Yanchun Liang,Chong Wang

doi:10.1109/access.2019.2949106

Abstract

It is always difficult to evacuate crowds in public places like subway stations. The traditional crowd behavior simulation models often ignore two important issues in crowd evacuation: pedestrian tracking and individual differences. To solve the problem, this paper combines social force model (SFM) with deep learning into a novel pedestrian detection method. Firstly, several deep learning algorithms for pedestrian detection were compared, and the best ones for sparse and dense crowds were determined. Next, the pedestrian positions in a real video were acquired by the selected algorithms, and converted into actual coordinates in the scene. Then, the evacuation process was simulated with our method and the SFM based on these coordinates. The results show that our model output closer-to-reality results than the SFM. The research findings shed important new light on evacuation in crowded areas.

Highlights

Many megacities have appeared across the globe for two intertwined reasons: the population boom and the accelerated urban growth
This paper extracts the actual position of pedestrians by the deep learning algorithm Faster region-based convolutional neural network (R-CNN)
For sparse and dense crowd, we selected the optimal pedestrian detection algorithm respectively and did the migration learning according to the dataset of this paper

Summary

Introduction

Many megacities have appeared across the globe for two intertwined reasons: the population boom and the accelerated urban growth. There are 36 cities with over 10 million residents around the world. The Chinese mainland boasts 15 cities with a population in excess of 10 million. With city’s development, the urban traffic environment is becoming more complex. Like subway stations and bus stops can be very crowded. People usually enter and exit these places in order. Accidents like crowding and trampling may happen in emergencies (e.g. fire or earthquake). It may lead to greater congestion, or more severe accidents, due to large gatherings of people at transport hubs. To design rational evacuation routes and strategies, many scholars have attempted to develop suitable models for real-world scenes and crowd behaviors [1]–[5]

Objectives

Methods

Findings

Conclusion