Terminal AI Model ID Selection for Low-Overhead Channel Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing radio communication technologies fail to adequately utilize artificial intelligence (AI) for overhead reduction and accurate channel estimation, resulting in insufficient effectiveness in existing technologies, which hinders communication throughput and quality improvements.
Innovation Solution
A terminal and base station system that utilizes AI models, identified by specific IDs such as meta IDs, to manage and control network operations, enabling efficient resource use and channel estimation through model compilation and information exchange.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If AI technology is utilized for network control and management, then communication throughput and quality can be improved, but overhead reduction and channel estimation accuracy cannot be achieved without sufficient studies
Solution Approach 1:
The model ID is segmented into multiple parts: a first ID indicating the AI model type and a second ID indicating the area or PLMN. This segmentation allows for efficient identification and selection of appropriate AI models without transmitting complete model information, thereby reducing overhead while enabling effective AI utilization for throughput improvement
Solution Approach 2:
AI models are pre-stored in terminals with their corresponding model IDs. The terminal can quickly determine and select the appropriate pre-stored model based on received ID information without needing to download or process complete model data, achieving both fast response for throughput improvement and reduced signaling overhead
2Reliability
If AI technology is utilized for network control and management, then communication quality can be improved, but channel estimation accuracy cannot be achieved without sufficient studies
Solution Approach 1:
The system implements feedback mechanisms where the terminal determines the appropriate AI model based on received ID information and communicates this selection back to the base station. This feedback loop ensures that the correct pre-stored model is used for channel estimation, improving both communication quality and estimation accuracy through iterative optimization
Solution Approach 2:
The invention changes the parameter representation from transmitting complete model data to transmitting compact ID parameters. The first ID (model type) and second ID (area/PLMN) serve as key parameters that enable the terminal to select the appropriate AI model, achieving accurate channel estimation while reducing signaling overhead
3Measurement precision
If complete AI model information is transmitted, then model selection accuracy can be improved, but signaling overhead increases
Solution Approach 1:
Instead of transmitting complete AI model information, the system transmits a simplified copy in the form of model IDs (first ID and second ID). The terminal uses these ID copies to identify and select from pre-stored complete models, maintaining model selection accuracy while dramatically reducing signaling overhead
Solution Approach 2:
The model ID structure serves multiple functions: the first ID identifies the AI model type, the second ID identifies the area or PLMN, and together they enable model selection, validation, and communication. This multi-functional ID system replaces the need to transmit detailed model information for each purpose separately, reducing overall overhead while maintaining accuracy
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A terminal according to one aspect of the present disclosure includes a receiving section that receives information related to an identifier (ID) of a model, and a control section that determines the model corresponding to the ID of the model, based on the information. The information includes the ID of the model and an ID related to an area or a Public Land Mobile Network (PLMN). According to one aspect of the present disclosure, preferable overhead reduction/channel estimation/resource use can be achieved.