AI-Based CSI Compression and Quantization for Accurate Reporting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems face challenges in efficiently compressing and transmitting large CSI measurement reports, leading to increased data size and reduced accuracy due to conventional linear compression methods.
Innovation Solution
Implementing AI/ML models at UE and network entities for compressing and quantizing CSI data, using scalar and codebook quantization to reduce data exchange, and reconstructing input data at the network entity, thereby reducing the amount of data transmitted.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional linear compression methods are used to compress CSI measurement reports, then the data size is reduced, but the accuracy of CSI reporting deteriorates
Solution Approach 1:
The patent changes the compression approach from linear compression to AI/ML-based compression, fundamentally altering the parameter transformation method. This allows for more intelligent compression that preserves accuracy while reducing data size, as the AI models learn optimal compression characteristics from training data.
Solution Approach 2:
The patent uses AI/ML models to create compressed representations (copies) of CSI data that maintain the essential information content. The models learn to create efficient representations during training and replicate this compression capability during operation, achieving both size reduction and accuracy preservation.
2Quantity of substance
If AI/ML models are implemented for compressing and quantizing CSI data, then the data exchange amount is reduced, but the device complexity increases
Solution Approach 1:
The patent segments the compression function into separate components: an AI/ML compression model and a quantization component. This segmentation allows the complex AI processing to be separated from the data transmission path, with only the compressed quantized data being transmitted, thus reducing data exchange while managing complexity through modular design.
Solution Approach 2:
The patent performs compression and quantization operations before data transmission. By pre-processing the CSI data through AI/ML models and quantization at the source (UE), the system reduces the amount of data that needs to be transmitted over the channel, thereby reducing data exchange amount and associated complexity in the transmission path.
3Measurement precision
If more data is transmitted to maintain CSI accuracy, then the measurement precision is improved, but the use of energy increases
Solution Approach 1:
The patent changes the data representation parameters through AI/ML-based compression and quantization. By transforming the CSI data into a compressed representation with fewer bits while preserving accuracy, the system reduces the energy required for transmission without sacrificing measurement precision.
Solution Approach 2:
The patent creates efficient data copies through AI/ML compression that maintain the essential information content with reduced data volume. These compressed copies require less energy to transmit while preserving the necessary accuracy for CSI reporting, thus resolving the contradiction between precision and energy consumption.
Data Source
AI summary
Various aspects of the present disclosure relate to methods, apparatuses, and systems that support artificial intelligence (AI) for channel state information (CSI). For instance, implementations provide an architecture and associated signaling for compressing an input (e.g., CSI at a user equipment (UE)), quantizing the compressed input, transmitting the quantized compressed input, and extracting (e.g., at a network entity such as a gNB) relevant information from the quantized compressed input. In at least some implementations the architecture includes one or more AI/machine learning (ML) models and is composed of multiple components, such as a UE component and a network entity component.