Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Distributed reinforcement learning social navigation method based on image hidden variable probability model

A probabilistic model and reinforcement learning technology, applied to instruments, adaptive control, control/regulation systems, etc.

Pending Publication Date: 2021-06-11
ZHEJIANG UNIV
View PDF8 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the training of this type of algorithm often requires a large amount of interactive data in the simulation environment, which is difficult to meet in the real environment or within a limited time.

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
  • Distributed reinforcement learning social navigation method based on image hidden variable probability model
  • Distributed reinforcement learning social navigation method based on image hidden variable probability model
  • Distributed reinforcement learning social navigation method based on image hidden variable probability model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0061] The present invention will be further elaborated and illustrated below in conjunction with the accompanying drawings and specific embodiments. The technical features of the various implementations in the present invention can be combined accordingly on the premise that there is no conflict with each other.

[0062] In a preferred embodiment of the present invention, a distributed reinforcement learning social navigation method based on image hidden variable probability model is provided,

[0063] Existing reinforcement learning social navigation methods learn desired social navigation policies through extensive trial-and-error for mobile robots in the environment. Since physical experiments will pose a threat to the safety of pedestrians, and the cost of physical robots is relatively high, this type of method generally chooses to perform training iterations in a simulation environment. However, the reinforcement learning algorithm still has shortcomings such as time-co...

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 distributed reinforcement learning social navigation method based on an image hidden variable probability model. According to the method, the image hidden variable probability prediction model is introduced to replace a traditional decisive prediction model, so that on one hand, the reasonability of the model is enhanced, prediction is closer to a pedestrian motion model with randomness, and on the other hand, the strategy performance is further enhanced by enhancing the exploration capability of the model, and overfitting is avoided. Meanwhile, the method achieves discrimination of a dynamic obstacle by decoupling movement of the mobile robot through an image sequence, a high-layer pedestrian detection module with instability is omitted, and sound migration is realized. In addition, a strategy sharing multi-agent simulation environment is designed to simulate a pedestrian dynamic environment, and the similarity between simulation and real crowd interaction is enhanced. In the environment, multiple agents synchronously perform data acquisition, so that the training time can be further shortened.

Description

technical field [0001] The invention belongs to the field of mobile robot navigation, and in particular relates to a distributed reinforcement learning social navigation method based on an image hidden variable probability model. Background technique [0002] With the rapid development of computer and automation technology, robot technology has gradually penetrated into all walks of life and into people's daily life. They not only assist or replace humans in complex and heavy work in traditional manufacturing, but also gradually replace humans in fields such as entertainment, medical care, and security. Under the trend of surging labor costs, the promotion of robot technology will undoubtedly greatly relieve the accompanying pressure. Therefore, robot technology is gradually becoming a current hot research direction, and service robot, as an important branch, has attracted countless attentions. [0003] As an important member of the robotics field, service robots are posit...

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
IPC IPC(8): G05B13/04
CPCG05B13/042
Inventor 熊蓉崔瑜翔王越
Owner ZHEJIANG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products