Abstract | ||
---|---|---|
•We propose a reinforcement learning based data-driven crowd evacuation (RL-DCE) framework for simulating actual human evacuation behaviors in dynamic environment.•A novel data-driven crowd evacuation (DCE) model is established, which can enhance the visual realism of crowd evacuation and the path calculation efficiency.•We propose a path planning method that is based on reinforcement learning and hierarchical mechanism, which allows individuals to learn the main observed features of crowd motion and demonstrate robustness with respect to dynamic environment.•We establish a simulation system based on the DCE model to visualize the analysis in a graphical manner. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1016/j.neucom.2019.08.021 | Neurocomputing |
Keywords | Field | DocType |
Reinforcement learning,Data-driven,Crowd evacuation,Cohesiveness | Motion planning,Crowds,Data-driven,Group cohesiveness,Collision,Robustness (computer science),Crowd simulation,Artificial intelligence,Mathematics,Machine learning,Reinforcement learning | Journal |
Volume | ISSN | Citations |
366 | 0925-2312 | 3 |
PageRank | References | Authors |
0.36 | 0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Zhenzhen Yao | 1 | 3 | 0.36 |
Guijuan Zhang | 2 | 67 | 13.28 |
Dianjie Lu | 3 | 52 | 10.88 |
Hong Liu | 4 | 139 | 22.83 |