A heterogeneous network multi-spectrum coordination method, device and medium

Through a multi-agent multi-spectrum collaborative approach, real-time reinforcement learning and neural networks are used to optimize frequency band resource allocation, solving the user experience issues of different service types under multi-band networking in heterogeneous networks, and achieving efficient multi-type service integration and resource scheduling.

CN119946884BActive Publication Date: 2025-10-21BEIJING UNIV OF TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510109943.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2024-02-04
Filing Date
2025-01-23
Publication Date
2025-10-21
Estimated Expiration
2045-01-23

AI Technical Summary

Technical Problem

Existing technologies fail to effectively address the latency, rate, reliability, and power consumption requirements of different service types in multi-band networking in heterogeneous networks, especially the inherent network intelligence requirements of Internet of Vehicles services and immersive communications, resulting in poor user experience.

Method used

It adopts a multi-agent multi-spectrum collaborative method, obtains user information and uses neural network units of real-time reinforcement learning, experience replay and target reinforcement learning to optimize frequency band resource allocation and user scheduling, thereby achieving seamless integration of multiple types of services.

Benefits of technology

It improves the user experience of different types of services in heterogeneous networks, meets various business needs through a lightweight multi-spectrum collaborative approach, reduces algorithm complexity and improves reinforcement learning efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119946884B_ABST
    Figure CN119946884B_ABST
Patent Text Reader

Abstract

The application discloses a heterogeneous network multi-frequency band coordination method, device and medium. The method comprises the following steps: inputting user measurement information into multiple agents of a multi-agent multi-frequency band coordination algorithm, executing a real-time reinforcement learning unit, outputting Q values to an experience replay unit, and triggering a timer 1; if the timer 1 does not stop, the experience replay unit continuously iterates; if the timer 1 stops, the result of the experience replay unit is output to a target reinforcement learning unit, the Q values are trained, and a timer 2 is triggered; if the timer 2 stops, the target reinforcement learning unit outputs the obtained Q values to the real-time reinforcement learning unit; if the timer 2 does not stop, the target reinforcement learning unit continuously iterates; and the multiple agents of the multi-agent multi-frequency band coordination algorithm output multiple Q values, and each Q value corresponds to a frequency band resource allocation result of one or a plurality of users. The application provides technical support for various types of services, communication terminals and communication networks.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • D2D communication network slice allocation method based on deep reinforcement learning

    CN113163451A

  • Wireless positioning device and method based on multi-band CSI cooperation

    CN113938823A