Dynamic Configuration for AI-Assisted Communication Reporting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication systems face challenges in efficiently reporting channel state information (CSI) and other AI and sensing-related information due to fixed configuration and reporting mechanisms, which cannot adapt to the dynamic and varying nature of these data.
Innovation Solution
A communication method and apparatus that allow a terminal device to obtain and send configuration information dynamically, including parameters such as quantization granularity, feedback periodicity, and resource allocation, to adapt to the specific requirements of the information being reported, thereby enhancing resource utilization and spectrum efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If fixed configuration information is used for CSI reporting, then the system structure is simple and easy to implement, but the system cannot adapt to dynamic information characteristics and has low resource utilization
Solution Approach 1:
The patent implements dynamic configuration information that can be adjusted based on information characteristics such as time-varying properties, importance, and size. The network device determines configuration parameters including feedback periodicity, resource allocation, and quantization granularity dynamically rather than using fixed configurations, allowing the system to adapt to changing conditions while maintaining manageable complexity through automated determination rules
Solution Approach 2:
The patent changes key configuration parameters such as feedback periodicity, resource allocation, and quantization granularity based on information characteristics. By adjusting these parameters dynamically according to the actual needs of different information types (e.g., AI training data vs. sensing data), the system achieves adaptability without requiring complete restructuring of the configuration framework
2Productivity
If fixed feedback periodicity is used, then the system is easy to manage, but resource utilization is low when information changes frequently or slowly
Solution Approach 1:
The patent dynamically adjusts feedback periodicity based on information characteristics such as time-varying properties and importance. For rapidly changing information like channel state, shorter periodicity is used, while for stable information like AI training data, longer periodicity is applied. This parameter adaptation improves resource utilization by matching feedback frequency to actual information dynamics
Solution Approach 2:
The system transitions from static feedback periodicity to dynamic periodicity determination. The network device automatically adjusts feedback timing based on real-time assessment of information characteristics, enabling the system to optimize resource usage for each information type while maintaining manageable complexity through automated decision rules
3Productivity
If uniform quantization granularity is used for all information, then the configuration is simple, but resource waste occurs for different information types with different precision requirements
Solution Approach 1:
The patent applies different quantization granularities to different information types based on their specific requirements. For example, AI training data may use coarser quantization while sensing data requires finer granularity. This localized optimization of quantization precision improves spectrum utilization by matching resource allocation to actual information needs without requiring complex overall restructuring
Solution Approach 2:
The system dynamically adjusts quantization granularity parameters according to information characteristics such as type, importance, and size. By changing this parameter adaptively for different information streams (e.g., using 2-bit quantization for some data and 4-bit for others), the system optimizes spectrum efficiency while maintaining manageable configuration complexity through automated determination
4Productivity
If multiple types of information are reported using the same mechanism, then the system structure is simple, but resource utilization is insufficient for diverse information characteristics
Solution Approach 1:
The patent segments the reporting mechanism into multiple independent configuration sets, each optimized for specific information types. Different configuration parameters (periodicity, resources, quantization) are determined separately for AI training data, sensing data, and other information types. This segmentation allows tailored optimization for each information category while maintaining overall system manageability through modular configuration
Data Source
AI summary
This application provides a communication method and apparatus. The method includes: A terminal device obtains first configuration information, where the first configuration information includes one or more of the following: a type of first information, a quantization granularity of the first information, a feedback resource of the first information, a feedback periodicity of the first information, a measurement resource corresponding to the first information, a reference signal corresponding to the first information, and a mapping relationship between the reference signal corresponding to the first information and the measurement resource corresponding to the first information. The terminal device sends the first information to a network device based on the first configuration information, where the first information is used to assist an artificial intelligence (AI) model of the network device. The network device assists the AI model of the network device based on the first information.