UE Location Estimation Using User Plane AI Model Transfer

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing 3GPP technologies face challenges in implementing AI-based UE location estimation, particularly in indoor factory scenarios, due to the large size of AI models exceeding the capacity of control plane messages and the need for efficient transfer and application of these models in user equipment.

Innovation Solution

The proposed solution involves transferring AI models for UE location estimation from the core network side using user plane protocols, enabling the download and application of AI models for fingerprinting-based positioning, which includes providing information on model validity, reference signals, and UE capabilities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If AI models are transferred using control plane messages, then the transfer can be managed through existing signaling protocols, but the large size of AI models exceeds the capacity of control plane messages

Engineering Contradiction:

ImproveAI model transfer reliabilityVSAvoidAI model size

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The patent introduces a user plane as an intermediary channel for AI model transfer. Instead of directly using control plane messages which have size limitations, the system uses user plane data channels (such as HTTP/2, FTP, or other IP-based data channels) as a mediator to handle the large-volume AI model transfer, thereby resolving the contradiction between reliable transfer and message size capacity

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transitions from a single-dimensional control plane transfer approach to a multi-dimensional approach by utilizing the user plane dimension. This allows AI models to be transferred through data channels that have much larger capacity, effectively adding a new dimension for model transfer that bypasses the size constraints of control plane messages

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Productivity

If AI models are downloaded from URL or FQDN via user plane protocol, then large AI models can be transferred efficiently, but additional infrastructure and configuration are required

Engineering Contradiction:

ImproveAI model transfer efficiencyVSAvoidmodel transfer infrastructure

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by having the LMF pre-configure URL or FQDN information in assistance data before the UE needs to download the AI model. This allows the UE to directly access the model through standard web protocols without requiring complex real-time negotiation or additional infrastructure setup during the positioning process

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The UE autonomously downloads the AI model using the provided URL or FQDN through standard user plane protocols. The device self-services by handling the model acquisition independently using existing internet infrastructure, eliminating the need for specialized dedicated transfer infrastructure

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20260046811A1Method and apparatus of user equipment (UE) location estimation

Publication Date: 2026.02.12 LENOVO (BEIJING) LTD

AI summary

Embodiments of the present application relate to a method and apparatus of user equipment (UE) location estimation. An exemplary method may include: sending a first message to an LMF associated with one or more BSs of a wireless network; receiving a second message from the LMF in response to sending the first message, wherein the second message at least includes information for determining an AI model for UE position estimation; and receiving the AI model as determined from the LMF.