Terminal AI Model Determination via Network Indication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication technologies face challenges in efficiently determining and utilizing AI models for beam management in terminals, leading to increased complexity and signaling overhead.
Innovation Solution
A method and device for determining a target AI model for use by a terminal, involving the reception of model indication information from an access network device and subsequent determination or configuration of the target model based on this information, utilizing various AI models such as machine learning and deep learning models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If AI models are used for beam management in terminals, then performance prediction capability is improved, but device complexity increases
Solution Approach 1:
The patent extracts the AI model determination and configuration functions from the terminal device and relocates them to the network side (access network device). The terminal only needs to receive model indication information and apply the indicated model, while the network device handles model selection, configuration, and management. This extraction significantly reduces terminal complexity while maintaining performance prediction capabilities.
Solution Approach 2:
The patent introduces model indication information as an intermediary mechanism between the network device and terminal. This intermediary carries essential model configuration parameters (such as model type, precision, input dimension) without requiring the terminal to independently determine or store multiple model options, thereby simplifying terminal operations while enabling sophisticated AI model management.
2Adaptability or versatility
If multiple AI models are supported for different scenarios, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent extracts the multi-model management functionality from the terminal and places it in the network device. The terminal simply receives model indication information specifying which model to use for current conditions, while the network device maintains the library of different AI models and determines the appropriate one based on scenario requirements. This enables high adaptability without increasing terminal complexity.
Solution Approach 2:
The patent creates a universal model indication information structure that can specify different model types (beam management models, channel prediction models, etc.) through a unified configuration mechanism. The same indication information framework handles various model scenarios, making the system adaptable to multiple AI applications without requiring separate terminal functionalities for each model type.
3Device complexity
If AI models are configured for beam management, then measurement complexity is reduced, but signaling overhead increases
Solution Approach 1:
The patent extracts complex model configuration details from repeated terminal measurements and relocates them to network-side model indication information. Instead of the terminal performing complex measurements to determine model parameters, the network device provides pre-determined model configurations through compact indication information, reducing measurement complexity while minimizing signaling overhead through efficient parameter encoding.
Solution Approach 2:
The patent transforms detailed model configuration parameters into compact model indication information that can be efficiently transmitted. By changing the representation from full model specifications to condensed indication parameters (model type identifiers, precision levels, input dimensions), the system reduces signaling overhead while still enabling precise AI model configuration for beam management.
Data Source
AI summary
Provided in the present application are a method and device for determining a model for use by a terminal. The method includes: receiving, by a terminal, model indication information from an access network device; and determining, by the terminal, a target model for use by the terminal based on the model indication information.