A crowd evacuation simulation method and system based on multi-agent deep reinforcement learning

A multi-agent, reinforcement learning technology, applied in the field of crowd evacuation simulation methods and systems, can solve problems such as unsatisfactory effects, crowded exits, and evacuation efficiency needs to be improved

CN109670270AInactive Publication Date: 2019-04-23SHANDONG NORMAL UNIV
3 Cites 42 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2019-04-23
Estimated Expiration
Not applicable · inactive patent

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a crowd evacuation simulation method and system based on multi-agent deep reinforcement learning. The method comprises the steps of creating a simulation scene according to aninitial coordinate and a motion speed of an individual in crowd evacuation; arranging a counter at each evacuation exit of the evacuation scene, calculating the congestion degree of the exit accordingto the area and the number of people, wherein the congestion degree is feedback of return rewards when a path is trained in the deep reinforcement learning model; grouping all individuals according to the position of each individual away from the exit of the room in each sub-region, and selecting the individual at the foremost end of the local region in the group as an in-group leader; using a multi-agent deep deterministic policy gradient algorithm MADDPG for path planning of leaders, regarding the multiple leaders as multiple agents, enabling the multiple agents to cooperate with one another to select an optimal evacuation path, and enabling the leaders to evacuate according to the path planned through deep reinforcement learning; enabling each member within the group follows the leaderto evacuate under improved social forces.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of multi-agent reinforcement learning and computer simulation, in particular to a crowd evacuation simulation method and system based on multi-agent deep reinforcement learning. Background technique

[0002] The statements in this section merely enhance the background related to the present disclosure and may not necessarily constitute prior art.

[0003] With the continuous acceleration of the urbanization process, the buildings and the density of people in the city are also increasing rapidly, followed by a large number of people gathering in public places, and in densely populated public places, because people are not familiar with the environment, Once an emergency occurs, it is very easy to cause vicious events such as crowd congestion and stampede. If the crowd cannot be evacuated effectively, it will often lead to vicious accidents such as mass death and mass injury. How to effectively carry out disa...

Examples

Embodiment Construction

[0078] It should be pointed out that the following detailed description is exemplary and is intended to provide further explanation to the present application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0079] It should be noted that the terminology used here is only for describing specific implementations, and is not intended to limit the exemplary implementations according to the present application. As used herein, unless the context clearly dictates otherwise, the singular is intended to include the plural, and it should also be understood that when the terms "comprising" and / or "comprising" are used in this specification, they mean There are features, steps, operations, means, components and / or combinations thereof.

[0080] Explanation of technical terms

[0081] KLT (tracking algorithm), English full name: Kanade-Lucas-Tomas...