Beam Pair Mapping for Low-Overhead Beam Training

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The beam training process in systems with large-scale antenna arrays is inefficient due to the high number of beams, leading to excessive overhead in beam alignment and adjustment.

Innovation Solution

A method involving the establishment of a mapping relationship between Beam Pair Link (BPL) identifiers, Channel State Information Reference Signal (CSI-RS) Resource Indicators (CRIs), and transmitting/receiving beams, allowing the transmitting end to configure reference signals with BPL identifiers, reducing the need for explicit beam information exchange.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If all transmitting beams are sent for beam training, then the receiving end can search for the best receiving beam, but the beam training process becomes expensive and inefficient due to the large number of beams

Engineering Contradiction:
Improvebeam alignment accuracyVSAvoidbeam training efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent segments the large set of transmitting beams into multiple beam groups. Instead of transmitting all beams individually for training, the system divides them into manageable groups and transmits reference signals for each group separately. This segmentation reduces the overhead of beam training while maintaining the ability to identify the best beam pair link through group-based comparison and selection.

Inventive Principle:
Principle #1Segmentation

2Reliability

If the number of antenna oscillators is increased to 512 or 1024, then the beam-forming gain is improved, but the system complexity and beam training overhead increase significantly

Engineering Contradiction:
Improvesignal coverageVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the large number of antenna oscillators into multiple antenna groups, where each group can be independently controlled for beamforming. This allows the system to manage the complexity of large-scale antenna arrays by processing beams in groups rather than individually, reducing the training overhead while maintaining the beam-forming gain benefits of having 512 or 1024 antenna oscillators.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements partial action by transmitting reference signals for only a subset of beams at each training stage rather than all beams simultaneously. The beam training process is performed in multiple stages with progressively refined beam selection, reducing the immediate complexity burden while still achieving comprehensive beam optimization through iterative refinement.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4050832B1Beam training method and device
Publication Date: 2026.03.04 DATANG MOBILE COMM EQUIP CO LTD
  • EP4050832B1 patent drawingFigure 1
  • EP4050832B1 patent drawingFigure 2
  • EP4050832B1 patent drawingFigure 3

AI summary

Disclosed are beam training methods and devices, used for reducing the overheads of a beam training process between a transmitting end and a receiving end. A beam training method includes: a receiving end determines a second mapping relationship; the receiving end receives configuration information of first channel state information reference signal resource transmitted by a transmitting end, the configuration information includes a first BPL identifier; the receiving end receives a first reference signal transmitted by the transmitting end using a first transmitting beam within the first channel state information reference signal resource; and the receiving end updates a first receiving beam corresponding to the first BPL identifier in the second mapping relationship, according to a measure result for the first reference signal.