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

Large-scale monitoring method and monitoring robot based on deep weighted double q learning

A robot, a large-scale technology, applied in neural learning methods, machine learning, program-controlled manipulators, etc., can solve the problems that the monitoring system cannot be fully deployed and the camera capacity is limited.

Active Publication Date: 2019-11-22
POLIXIR TECH LTD
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the invention is to provide a large-scale monitoring method based on deep weighted double-Q learning, which not only solves the problem that the monitoring system cannot be fully deployed due to the large monitoring range, but also solves the problem of limited camera capacity

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
  • Large-scale monitoring method and monitoring robot based on deep weighted double q learning
  • Large-scale monitoring method and monitoring robot based on deep weighted double q learning
  • Large-scale monitoring method and monitoring robot based on deep weighted double q learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0071] see figure 1 , as shown in the legend, is a wide-range monitoring robot 10 based on deep weighted double-Q learning. The Q-value table of the robot 10 includes Q A Table and Q B Table, the Q value is calculated by the depth estimation network parameter θ, where,

[0072] Q A The update formula for the value is as follows:

[0073]

[0074] δ=R(s,a)+γ[β A Q A (s',a * ;θ)+(1-β A ) Q B (s',a * ;θ)]-Q A (s, a; θ);

[0075] Q A ← Q A (s,a;θ)+α(s,a)δ;

[0076] Q B The update formula for the value is as follows:

[0077]

[0078] δ=R(s,a)+γ[β B Q B (s',a * ;θ)+(1-β B ) Q A (s',a * ;θ)]-Q B (s, a; θ);

[0079] Q B ← Q B (s,a;θ)+α(s,a)δ;

[0080] Among them, β A , β B Indicates the weight; s' indicates the next state; a * Indicates the optimal action for the next state; a L Represents the worst action in the next state; c is a free parameter, c≥0; δ represents time difference; R represents reward value; γ represents target discount, 0≤γ≤1; s r...

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 large-scale monitoring method based on deep weighted double Q learning. First, a Q value table including Q A Table and Q B The robot in the table, secondly, the unknown object enters a large-scale space to trigger the robot, thirdly, the robot perceives the current state s, and judges whether the current state s is the target state, if so, the robot reaches the next state and monitors the unknown object, if not, the robot To the next state, the robot gets the reward value according to the next state, and the robot chooses with equal probability to update Q A value or Q B value, and then update the Q value until an optimal monitoring strategy is obtained through convergence. The invention not only solves the problems of limited monitoring range and limited camera capacity, but also does not need to consider the synchronization of multiple cameras, thereby reducing the cost. The invention also discloses a large-scale monitoring robot based on deep weighted double Q learning.

Description

technical field [0001] The invention relates to a large-scale monitoring field, in particular to a large-scale monitoring method and a monitoring robot based on deep weighted double-Q learning. Background technique [0002] In our daily life, monitoring systems are ubiquitous, such as traffic light monitoring at traffic intersections, security monitoring in residential areas, etc. The monitoring system combines multimedia technology, computer network, industrial control and artificial intelligence and other aspects of knowledge, which can be used for security prevention, information acquisition, dispatching and commanding, and can also provide various services for production processes and distance education. However, in some large-scale environments that need to complete specific tasks, such as finding and tracking and monitoring unknown objects, the current monitoring system cannot be fully deployed. The reason is: on the one hand, due to the inherent defects of surveillan...

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 Patents(China)
IPC IPC(8): G06N3/04G06N3/08H04N7/18
CPCG06N20/00H04N7/18B25J9/163G05B2219/40264G05B2219/40298B25J9/1697G05B13/027
Inventor 章宗长潘致远王辉
Owner POLIXIR TECH LTD
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com