BR-102024017502-A2 - Multi-Agent Adaptive Beam Scanning Framework for 6G Systems

Abstract

The present invention incorporates adaptive multi-agent algorithms into a beam scanning framework to select the most suitable beam set to transmit reference signals within a predetermined time interval to cover various angular regions of interest. Multiple transmitting nodes undergo an adaptive learning process and cooperate with each other to share available information and acquired knowledge to determine the most appropriate beam subset from a large set of available beams to transmit reference signals through radio channels to multiple receiving nodes. The transmitting nodes are endowed with intelligence and establish cooperation strategies to expedite their learning process; that is, they acquire knowledge from their interaction with other transmitting and receiving nodes in the wireless network to perform beam scanning with reduced signaling overhead, latency, and power consumption. The present invention improves beam management in extremely dense networks with higher carrier frequencies and provides efficient signal exchange seeking scalability and reliability with low latency, making a significant contribution to the development of sixth-generation (6G) mobile communication networks.

Inventors

• FRANCISCO HUGO COSTA NETO
• DERCIO MANUEL MATE
• DANIEL DA SILVA SOUZA
• THAIS CARVALHO AREIAS
• MARIO OLIVEIRA COSTA DIAS

Assignees

• Samsung Eletrônica da Amazônia Ltda.

Dates

Publication Date

20260310

Application Date

20240826

Claims (11)

1. 1. A method for adaptive beam selection for multiple network nodes, characterized in that it comprises: initializing the policies of all N intelligent nodes according to the network operator's objectives; initializing the mapping between the current beams transmitting reference signals (states), the set of selected beams (actions), and the calculated values of the cost (reward) functions for each intelligent node; determining a set of beams to transmit reference signals in a burst action set. where the parameter T corresponds to the total number of reference signals; transmit a set of reference signals contained in A to associated nodes; receive measurements of the signal level sent by the network and calculate the associated cost-reward function (R), which comprises a set of associated weights. ,where the parameter X corresponds to the number of network parameters considered in the cost function; update the mapping of the burst action set A with its associated reward cost function (R); retransmit an updated burst action set A'; and update the policies of all intelligently equipped nodes.
2. 2. Method, according to claim 1, characterized in that the determination of the beamset also includes considering signal level measurements and quality indication.
3. 3. A method according to claim 1, characterized in that the network node executes actions iteratively and measures the impact of the decisions made on the network's performance.
4. 4. Method, according to claim 1, characterized in that the weights are updated according to the network objectives.
5. 5. Method, according to claim 4, characterized in that the network performance is based on signal level measurements or quality indication.
6. 6. A method according to claim 1, characterized in that if a negative impact on network performance is identified due to decisions made by the network node, a fallback is triggered to a predefined mechanism.
7. 7. Method according to claim 1, characterized in that the t-th element of the burst set A is limited to a set of beams , determined according to the angular region to be covered in the beam scan, where the parameter M corresponds to the number of beams.
8. 8. Method, according to claim 1, characterized in that the parameters T and M are predefined by the network.
9. 9. A method according to claim 1, characterized in that the cost-reward function (R) is: where {Fi, F2 FX} are the network performance indicators and/or network objectives that quantify the effects of beam selection for transmitting SSBs in a burst set, and W = {W1, W2, ..., WX} is the set of associated weights.
10. 10. Method according to claim 1, characterized in that the reference signal is a synchronization signal block.
11. 11. A method according to claim 1, characterized in that the central controller updates the network node policies to accelerate beam scanning based on the correlation between beams from different network nodes.

Description

TECHNICAL FIELD [001] The present invention relates to the field of mobile communications network technology and relates to the method and apparatus for incorporating adaptive multi-agent algorithms into a beam scanning framework to select the appropriate set of beams to be employed in transmitting reference signals in scenarios with a large number of transmitters and receivers. [002] The fifth-generation (5G) mobile communications air interface has a beam-centric design. That is, channels and signals can be beamformed for data transmission and control plane capabilities. Therefore, 5G provides mechanisms to transmitting and receiving network nodes to establish highly directional communication links and exploit the benefits of the resulting beamforming gain. [003] These mechanisms are known as beam management and encompass various control tasks. The current beam management framework relies on beam scanning to determine the most suitable transmitter-receiver beam pairs in a given angular region. During beam scanning, each transmitter node sequentially sends reference signals through beams from an entire codebook or a subset of a codebook to perform quality measurements and use this information for beam indication. [004] The upcoming 3GPP releases (Release 18 onwards) will extend the frequency in the millimeter wave band from 52.6 GHz to 71 GHz. Furthermore, the sub-THz band from 92 GHz to 300 GHz is considered a new radio spectrum frontier for 6G development. That is, the exploration of higher carrier frequencies is a technological trend and will require even narrower beams and more antenna elements. Thus, the exploration of wider bandwidths will represent a significant increase in codebook size. [005] 5G was designed to support massive machine-like communications (mMTC) to promote a digital transformation of society and improve the overall efficiency of different vertical sectors. However, some of the most demanding requirements, such as reliability and latency for continuous operation of emerging mMTC applications, are still a concern. [006] 6G was designed to handle ever-increasing traffic for a growing number of devices. Therefore, operational changes in ultra-dense networks operating at higher frequencies will directly impact existing beam scanning procedures, since the sequential transmission-reception of reference signals through large codebook beams will be severely limited due to a large signaling overhead and higher latency. [007] The present invention incorporates adaptive multi-agent strategies in the beam scanning framework. Multiple transmitting nodes cooperate to optimize the selection of appropriate subsets of large codebook beams. These beams are employed in the transmission of reference signals in high-frequency operating scenarios with a large number of antenna elements and diverse users, for example, crowded massive MIMO scenarios in millimeter waves and THz. [008] The framework proposed in the present invention exploits a cooperative adaptive strategy between multiple transmitters to determine the appropriate beams of large codebooks to transmit reference signals with reduced signaling overhead and latency. [009] Several transmitters share available information with each other to accelerate the decision-making process and improve strategic beam selection. The solution proposed in the present invention offers great operational flexibility, as it can be defined in terms of one or more network parameters, such as transmission-reception beam link quality, latency, signaling, power, etc. [0010] The present invention explores adaptive cooperation between transmitting nodes to greatly reduce signaling overhead and latency due to search space optimization. The proposed idea is very flexible and adapts quickly to network changes, as it leverages the available information from each transmitter and shares it with the network to update decision-making strategies. BACKGROUND OF THE INVENTION [0011] The overall beam management framework was introduced in 3GPP Release 15. It encompasses several physical layer (PHY) and data link layer procedures (also known as L1 and L2 procedures, respectively) to acquire and maintain directional links between transmitter and receiver nodes. In this context, initial access leverages the functions of the physical and medium access control (MAC) layers to control directional communications and establish precise alignment of transmitter and receiver beams. Its main operations are beam scanning, beam measurement, beam determination, and beam recovery. [0012] Beam scanning refers to the process in which a transmitting node, for example, a base station (BS), user equipment (UE), or unmanned aerial vehicle (UAV), covers a spatial area by sequentially transmitting reference signals through different analog beams according to a predetermined sequence. During beam scanning, the transmitting node sequentially uses beams from an entire codebook or a subset of a codebook to find t