AI Protocol Layer Decoupling for Flexible Wireless Network Training

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The flexibility and accuracy of AI training in wireless networks are limited due to the binding of AI functions to existing protocol layer procedures, leading to insufficient and single-type data acquisition, which affects network performance and efficiency.

Innovation Solution

Introduce an independent AI protocol layer above the existing protocol layer to allow flexible adjustment of AI functions and data acquisition, enabling separate AI and communication functions, and using AI parameters to specify data type, source, and reporting conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If AI function is completed through existing NGAP protocol layer procedure, then communication function and AI function are integrated in same protocol layer, but flexibility of AI function is poor and AI function is bound to existing protocol layer procedure

Engineering Contradiction:
Improveflexibility of AI functionVSAvoidprotocol layer structure
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the protocol layer structure by introducing an independent AI protocol layer above the existing NGAP protocol layer. This segmentation separates AI functions from communication functions, allowing AI parameters to be flexibly configured and adjusted without modifying the underlying communication protocol structure. The AI protocol layer can independently define AI-specific procedures while reusing existing communication mechanisms.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts AI functions from the existing NGAP protocol layer by creating a dedicated AI protocol layer. This extraction allows AI parameters to be independently managed and configured, freeing them from the constraints of the original communication protocol structure while maintaining the integrity of both layers.

Inventive Principle:
Principle #2Taking out (Extraction)

2Quantity of substance

If MDT parameter is used for AI data acquisition, then existing protocol procedures are utilized, but data type is single and AI training requirements cannot be met

Engineering Contradiction:
Improvediversity of AI dataVSAvoidAI training efficiency
Core Design Contradiction:
Quantity of substanceVSProductivity

Solution Approach 1:

The patent applies local quality by defining AI-specific parameters within the AI protocol layer that are tailored to different AI training requirements. Instead of using generic MDT parameters, the system can configure specific AI parameters for different data types, collection methods, and reporting formats, allowing diverse data acquisition strategies to be implemented locally for different AI training scenarios.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent introduces dynamic configurability through AI parameters that can be flexibly adjusted based on specific AI training needs. The AI protocol layer enables dynamic modification of data collection requirements, reporting formats, and parameter configurations without being constrained by the static structure of existing MDT procedures, allowing the system to adapt to varying AI training demands.

Inventive Principle:
Principle #15Dynamics

3Reliability

If AI parameter is flexibly configured to meet various AI training requirements, then data acquisition can be optimized, but existing protocol layer may need modification

Engineering Contradiction:
Improveaccuracy of AI trainingVSAvoidimplementation complexity
Core Design Contradiction:
ReliabilityVSEase of manufacture

Solution Approach 1:

The patent implements a nested structure where the AI protocol layer is positioned above and contains AI-specific parameters and procedures, which in turn can utilize and nest existing NGAP protocol procedures for actual data transmission. This nested architecture allows flexible AI parameter configuration while relying on the established, stable underlying communication protocol, avoiding the need to modify existing protocols while achieving high AI training accuracy.

Inventive Principle:
Principle #7Nested doll (Nesting)

Data Source

PatentEP4090064B1Communication method and apparatus
Publication Date: 2025.07.30 HUAWEI TECH CO LTD
  • EP4090064B1 patent drawingFigure 1~2
  • EP4090064B1 patent drawingFigure 3~4
  • EP4090064B1 patent drawingFigure 5~6

AI summary

This application provides a communication method and apparatus. An independent artificial intelligence AI protocol layer is introduced to decouple an AI function from an existing next generation application protocol NGAP protocol layer procedure. The method includes: A first artificial intelligence AI protocol layer of a core network element generates an AI parameter, and sends the AI parameter to a first access network device. Then, the core network element receives, from the first access network device, AI data obtained based on the AI parameter, to complete network optimization. The first AI protocol layer of the core network element is an upper layer of a next generation application protocol NGAP protocol layer of the core network element, a first AI protocol layer of the first access network device is an upper layer of a next generation application protocol NGAP protocol layer of the first access network device, and the AI parameter is for indicating the AI data that needs to be obtained and an obtaining manner of the AI data.