Multi-Beam Positioning Model Training With UE Statistical Reporting

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional UE-assisted and UE-based positioning processes in wireless communication systems incur high network traffic, resource usage, and security concerns due to the transmission of raw measurement data, and limit the LMF's ability to determine UE positions.

Innovation Solution

Implementing a system where the LMF requests UEs to generate statistical values or position data based on specific reporting conditions, training machine learning models using these values, and distributing trained models to nearby UEs for position determination.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If raw measurement data is transmitted from UEs to LMF for positioning, then positioning accuracy is improved, but network traffic and resource consumption increase

Engineering Contradiction:

Improvepositioning accuracyVSAvoidnetwork traffic

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent extracts only the essential positioning information from raw measurement data by having UEs compute statistical values (mean, variance) and triggered data based on reporting conditions. This extraction occurs at the UE level before transmission to LMF, significantly reducing the quantity of data transmitted while preserving positioning accuracy.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent implements preliminary processing of measurement data at the UE level by computing statistical values and determining triggered data before transmission. This preliminary action reduces the data volume that needs to be transmitted over the network while maintaining the information needed for accurate positioning.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If raw measurement data is transmitted from UEs to LMF, then positioning capability is improved, but security concerns increase due to privacy

Engineering Contradiction:

Improvepositioning capabilityVSAvoidprivacy security

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent extracts only necessary positioning information from raw measurement data by computing statistical values and triggered data at the UE level. This extraction reduces the amount of personal information transmitted to LMF, thereby enhancing privacy security while maintaining positioning capability.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces statistical values and triggered data as intermediary representations of raw measurement data. These intermediaries convey positioning information without exposing raw personal data, acting as a mediator between UE and LMF that preserves both positioning capability and privacy security.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Measurement precision

If LMF aggregates measurement information from multiple UEs to train machine learning models, then model accuracy is improved, but data processing complexity increases

Engineering Contradiction:

Improvemodel accuracyVSAvoiddata processing complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary computation of statistical values and determination of triggered data at the UE level before data aggregation. This preliminary action simplifies the data format and structure before it reaches LMF, reducing the processing complexity required for aggregation and model training while maintaining model accuracy.

Inventive Principle:

Principle #10Preliminary action

4Quantity of substance

If statistical values and triggered data are transmitted instead of raw data, then data transmission is reduced, but positioning accuracy may be compromised

Engineering Contradiction:

Improvedata transmissionVSAvoidpositioning accuracy

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent transforms raw measurement data into statistical parameters (mean, variance) and triggered data representations. This parameter transformation reduces data transmission volume while preserving the essential information needed for accurate positioning, as the statistical parameters capture the key characteristics of the measurement data.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20260059491A1Apparatus and methods for machine learning model training in multi-beam communication systems

Publication Date: 2026.02.26 QUALCOMM INC

AI summary

Methods, systems, and apparatuses for training machine learning processes in multi-beam wireless communication systems. For example, a computing device generates a measurement request message for one or more statistical values that are determined based on beam measurements taken over a measurement interval. The computing device transmits the measurement request message to a user equipment, where the measurement request message causes the user equipment to determine the one or more statistical values based one or more beam measurements determined over one or more of the measurement intervals. Further, the computing device receives, from the user equipment, a measurement response message that includes the one or more statistical values. The computing device also trains a machine learning model based on the one or more statistical values.