AI/ML CSI Report Processing Unit Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The traditional PMI feedback method in wireless communication systems is inefficient due to high overhead, and the introduction of AI/ML-based CSI report technology requires different processing capabilities and resource demands compared to traditional methods.
Innovation Solution
A method for determining the number of processing units needed for AI/ML-based CSI reports by associating a CSI report configuration with an index, where the number of processing units occupied by the CSI report is related to the index, optimizing the assignment of processing units based on the specific demands of different CSI report configurations and radio channels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional PMI feedback method is used, then the system maintains compatibility with existing protocols, but the overhead becomes excessively high
Solution Approach 1:
The patent extracts only the essential CSI information needed for network operation and transmits it efficiently. By using AI/ML-based CSI calculation, the system extracts critical channel state information while reducing the amount of feedback data compared to traditional PMI methods, thus lowering overhead while maintaining protocol compatibility
Solution Approach 2:
The patent changes the fundamental parameters of CSI reporting by transitioning from traditional codebook-based PMI feedback to AI/ML-based CSI calculation. This parameter change enables more efficient information representation and reduces feedback overhead while maintaining system compatibility through standardized interfaces
2Productivity
If AI/ML-based CSI report technology is introduced, then the processing efficiency is improved, but the demand for processing units increases
Solution Approach 1:
The patent dynamically allocates processing units based on the specific demands of different CSI report configurations. The network equipment determines the number of processing units required by the UE for AI/ML-based CSI calculation and assigns resources accordingly, allowing the system to scale processing capacity dynamically rather than allocating fixed resources
Solution Approach 2:
The patent segments the processing unit requirements into different categories based on CSI report configuration types. By associating different numbers of processing units with different configuration indices, the system can allocate resources in discrete units matching the actual computational demands of various AI/ML algorithms
3Device complexity
If a unified solution is adopted for different CSI scenarios, then the hardware complexity is reduced, but the optimization capability for specific scenarios is limited
Solution Approach 1:
The patent creates a universal processing unit allocation mechanism that works across different CSI reporting scenarios including both traditional codebook-based and AI/ML-based methods. The network equipment uses a unified approach of associating configuration indices with processing unit requirements, allowing the same hardware infrastructure to support multiple scenarios without scenario-specific hardware additions
Data Source
AI summary
The present application provides a method and device in a node for wireless communications. A first node receives a first CSI (Channel Status Information) report configuration set; transmits a first information block. The first CSI report configuration set comprises a first CSI report configuration, the first CSI report configuration is used to determine a first CSI report, and the first CSI report comprises a first compress CSI; the first information block comprises the first CSI report; the first CSI report configuration is associated with a first index; the first CSI report occupies a first-type processing unit starting from a first symbol, and a number of the first-type processing unit(s) occupied by the first CSI report is related to the first index. The above method satisfies the demands for processing capability for AI/ML-based CSI report, while avoiding the waste of processing capability and optimizing the system configuration.