Method for using graph convolution reinforcement learning to minimize information age in group perception

A group perception and reinforcement learning technology, applied in the field of group perception, can solve the problems of uncontrollable movement of mobile users, difficult evaluation indicators, large energy consumption, etc., to minimize the age of information, avoid high-cost failures, and reduce excessive dependence. Effect

Pending Publication Date: 2022-01-14
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] 1. The perception cluster scheduling algorithm should achieve good results on multiple evaluation indicators at the same time, including collecting the data of each mobile user as soon as possible and efficiently, ensuring high user coverage, and saving energy consumption of the perception body as much as possible. However, the challenge lies in these It is difficult to balance the evaluation indicators. In order to improve the user coverage, the sensing body inevitably needs to collect data in relatively remote areas, resulting in a large energy consumption
[0006] 2. The movement of mobile users is random and uncontrollable, and the perception cluster needs to carefully design its trajectory according to the location distribution of mobile users over time, which is much more difficult than the task of scheduling sensory bodies to collect data from stationary objects

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
  • Method for using graph convolution reinforcement learning to minimize information age in group perception
  • Method for using graph convolution reinforcement learning to minimize information age in group perception
  • Method for using graph convolution reinforcement learning to minimize information age in group perception

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0069] The content of the present invention will be further described in detail below in conjunction with the accompanying drawings of the description. The method of the present invention comprises the following steps:

[0070] Step 1, such as figure 1 As shown, the perception platform opens the main process, establishes an empty experience multiplexing pool and initializes the parameters of the GCRL-min (AoI) algorithm, and the parameters of the algorithm include the parameters of the relational graph convolutional network, the parameters of the next state prediction module, the estimated the parameters of the value network;

[0071]Step 2. The sensing platform starts a sub-process that interacts with the environment. The sub-process interacts with the environment for one round, and establishes a simulation environment for a group sensing scene. U sensory bodies are deployed as executors for sensing data collection. There are M mobile devices in the environment. The initial...

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 method for using graph convolution reinforcement learning to minimize information age in group perception. The method comprises the following steps that: step 1, a sensing platform opens a main process; step 2, the sensing platform opens a sub-process interacting with the environment; step 3, states of a mobile user and a sensing body in a sensing area are observed, wherein the states comprise positions and moving directions of the sensing body and the user; step 4, interaction characteristics obtained by using a relational graph convolutional network are used as input; step 5, the step 3 and the step 4 are repeatedly executed until rounds are finished, step 6, batch empirical data is sampled from the empirical multiplexing pool by the host process of the sensing platform, and step 7, the step 2, the step 5 and the step 6 are repeatedly executed until the number of interaction rounds reaches an upper limit; and step 8, taking out the stored optimal parameters by the host process of the sensing platform. The method has the beneficial effects that the information age can be minimized, and excessive dependence on a simulation platform can be reduced.

Description

technical field [0001] The invention belongs to the technical field of group sensing, and in particular relates to a method for minimizing information age by using graph convolution reinforcement learning in group sensing. Background technique [0002] Crowd sensing has been recognized as an efficient and scalable way to obtain data for various smart city applications, such as traffic control and road condition monitoring. In the scenario of group sensing tasks, mobile users on the ground are constantly moving and obtaining data in the city. The sensing cluster (composed of multiple sensory bodies) serves as a mobile base station to serve mobile users and collect data obtained by mobile users for subsequent use. analysis processing. [0003] In group sensing tasks, many data have strong timeliness (such as traffic information), and the delay in obtaining real-time data from mobile users is a key indicator to measure the quality of task completion. How to design the trajecto...

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): G06F17/10G06F9/50G06N3/04G06N3/08
CPCG06F17/10G06F9/5027G06N3/08G06N3/045
Inventor 戴子彭刘驰叶语霄
Owner BEIJING INSTITUTE OF TECHNOLOGYGY
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