Beamforming in massive multiple-input multiple-output environment
The TD3-INVASE ML model optimizes beamforming in mMIMO systems by discovering causal relationships, addressing high training overhead and enabling efficient, generalized beamforming in dynamic environments.
WO2025220020A1PCT designated stage Publication Date: 2025-10-23TELEFONAKTIEBOLAGET LM ERICSSON (PUBL) +1
Patent Information
- Application Number
- PCT/IN2024/050398
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-04-15
- Publication Date
- 2025-10-23
Smart Images
Figure IN2024050398_23102025_PF_FP_ABST
Abstract
A computer-implemented method is provided that is performed by a computing device in a first environment comprising a plurality of mMIMO nodes. The method includes training (1004) a ML model to learn antenna beam patterns of at least one mMIMO node based on an identification of a relationship between an action space for the first environment and rewards for the ML model based on respective actions from the action space. The method further includes extracting (1006) a subset of the action space based on the relationship that include respective actions that maximize a beamforming gain of a plurality of antennas of the at least one mMIMO mode. Related methods and apparatus are also provided.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Patent Citations
Method and apparatus for machine learning based wide beam optimization in cellular network
WO2019231289A1
Reinforcement learning of beam codebooks for millimeter wave and terahertz MIMO systems
WO2023287769A2