Labyrinth navigation method and device based on multi-agent layered reinforcement learning

A reinforcement learning and multi-agent technology, applied in navigation, machine learning, measuring devices, etc., to achieve the effect of accelerating the convergence speed and slowing down the impact

Active Publication Date: 2021-08-06
TIANJIN UNIV
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The invention provides a maze navigation method and device based on multi-agent layered reinforcement learning. Environmental non-static problems caused by inconsistency, see the description below for details:

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Labyrinth navigation method and device based on multi-agent layered reinforcement learning
  • Labyrinth navigation method and device based on multi-agent layered reinforcement learning
  • Labyrinth navigation method and device based on multi-agent layered reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0030] In order to make the purpose, technical solution and advantages of the present invention clearer, the implementation manners of the present invention will be further described in detail below.

[0031] The embodiment of the present invention provides a model-based multi-agent layered reinforcement learning maze navigation method, see figure 1 and figure 2 , the method includes the following steps:

[0032] Step (1): Obtain the location information of the agent, initialize the parameters of each agent, and establish an initial maze environment model;

[0033] Step (2): Each agent uses a hierarchical structure to perform exploration actions, and judges whether there are obstacles around the agent. If there are obstacles, it performs obstacle avoidance actions, otherwise it performs navigation actions. After a period of exploration, the agents gradually reduce the use of Hierarchical selection action;

[0034] Step (3): The agent performs actions in the current state o...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a labyrinth navigation method and device based on multi-agent layered reinforcement learning, and the method comprises the steps: enabling each agent to execute an exploration action through a layered structure, judging whether there is an obstacle around the agent, executing an obstacle avoidance action if there is an obstacle, and enabling the agents to gradually reduce the selection action of the layered structure; the agent executes actions in the current labyrinth environment state, collects empirical data, updates an environment model according to the empirical data, judges whether the agent reaches a target point or collides with an obstacle or not, if yes, the agent begins to explore from the initial position again, and if not, the agent continues to explore in the labyrinth environment; the agents are trained by using the empirical data and the environmental model, and each agent bypasses an obstacle in the labyrinth environment and reaches a respective specified target point in a shortest path. The device comprises a processor and a memory. According to the method, the agents are enabled to find the target point more quickly, the number of interaction times is reduced, and the environment non-static problem caused by discordance between the agents is solved.

Description

technical field [0001] The invention relates to the field of multi-agent reinforcement learning, in particular to a maze navigation method and device based on multi-agent layered reinforcement learning. Background technique [0002] A multi-agent system is a group system composed of multiple autonomous individuals. Through communication, cooperation and competition among agents, it can complete complex tasks that a single agent cannot complete. Multi-agent maze navigation enables multiple agents to quickly navigate to a designated target point in a maze environment and complete specific tasks. It has been widely used in material transportation, fire rescue, field search and rescue, and warehouse logistics transportation. Therefore, the research on multi-agent maze navigation is of great significance. [0003] Reinforcement learning is a subfield of machine learning that is mainly used to deal with sequential decision problems. Reinforcement learning consists of two element...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G01C21/20G01C21/00G06N20/00
CPCG01C21/20G01C21/005G06N20/00
Inventor 穆朝絮刘朝阳朱鹏飞
Owner TIANJIN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap