Beamformed Reference Signal Measurement for Low-Overhead Data Collection

Solution Overview

Problem

Current 3GPP systems face challenges in data collection for AI/ML models due to high overhead and battery consumption, especially in beam management procedures, as legacy CSI acquisition methods are costly and not designed for data collection, particularly when many UEs need to perform measurements.

Innovation Solution

A method where observing UEs are indicated to perform measurements on downlink reference signals intended for another UE using a group PDCCH message with a data_collection_measurement-RNTI, allowing data collection without immediate reporting, thereby reducing overhead and conserving battery.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If legacy CSI acquisition methods are used for data collection, then measurement data can be obtained, but overhead and battery consumption increase significantly

Engineering Contradiction:

Improvedata collection capabilityVSAvoidbattery consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies universality by enabling UEs to perform measurements for multiple purposes simultaneously. The same downlink reference signals intended for one UE can be measured by multiple observing UEs for AI/ML model training, allowing the system to collect training data without dedicating separate resources. This multi-functional use of existing signals reduces overhead and energy consumption while enabling data collection.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent implements self-service by allowing observing UEs to autonomously perform measurements on downlink reference signals without requiring explicit triggering or coordination for each measurement opportunity. The UEs can independently collect measurement data for AI/ML training purposes, reducing the signaling overhead and energy consumption associated with centralized measurement scheduling.

Inventive Principle:

Principle #25Self-service

2Measurement precision

If legacy CSI acquisition methods are used for data collection, then measurement data can be obtained, but signaling overhead increases

Engineering Contradiction:

Improvedata collection capabilityVSAvoidoverhead

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent applies universality by enabling UEs to perform measurements for multiple purposes simultaneously. The same downlink reference signals intended for one UE can be measured by multiple observing UEs for AI/ML model training, allowing the system to collect training data without dedicating separate resources. This multi-functional use of existing signals reduces overhead and energy consumption while enabling data collection.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent implements self-service by allowing observing UEs to autonomously perform measurements on downlink reference signals without requiring explicit triggering or coordination for each measurement opportunity. The UEs can independently collect measurement data for AI/ML training purposes, reducing the signaling overhead and energy consumption associated with centralized measurement scheduling.

Inventive Principle:

Principle #25Self-service

3Quantity of substance

If many UEs perform measurements using legacy methods, then sufficient training data can be collected, but system overhead and energy consumption increase

Engineering Contradiction:

Improvetraining data volumeVSAvoiddata collection efficiency

Core Design Contradiction:

Quantity of substanceVSProductivity

Solution Approach 1:

The patent applies merging by combining the measurement activities of multiple observing UEs with the existing downlink reference signal transmissions. Instead of each UE requiring separate measurement resources, multiple UEs simultaneously observe the same reference signals intended for other UEs. This consolidation allows the system to collect training data from multiple UEs without proportionally increasing overhead or energy consumption, thereby improving data collection efficiency.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent applies universality by enabling UEs to perform measurements for multiple purposes simultaneously. The same downlink reference signals intended for one UE can be measured by multiple observing UEs for AI/ML model training, allowing the system to collect training data without dedicating separate resources. This multi-functional use of existing signals reduces overhead and energy consumption while enabling data collection.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20260040293A1Systems and methods for data collection for beamformed systems

Publication Date: 2026.02.05 TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

AI summary

A method (900) by a first user equipment, UE (108, 110), for data collection includes receiving (902), from a network node (104), a first message configuring one or more downlink reference signal resources. The UE receives (904), from the network node, a second message comprising an indication to monitor a communication channel associated with a radio network temporary identifier, RNTI, for data collection. The UE receives (906) a third message comprising downlink control information, DCI, associated with the RNTI for data collection. The third message indicates to the UE to perform measurements on one or more of the downlink reference signal resources. Based on the third message comprising the DCI associated with the RNTI for data collection, the UE performs (908) at least one measurement on a downlink reference signal transmitted for a second UE (106). The downlink reference signal is transmitted on one of the one or more indicated downlink reference signal resources.