Interpolation Neural Network for Channel State Information Recreation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CSI reporting in cell-free massive MIMO systems, especially with FDD systems at FR1 below 6 GHz, faces challenges in accurately reporting a multitude of relevant channel components, leading to high overhead and resource constraints, particularly in scarce UL resources and power consumption.
Innovation Solution
The method involves using an interpolation neural network trained with prior channel information to estimate channels between a terminal and a base station, allowing for the recreation of a high number of channel components with low overhead by leveraging learned prior knowledge and eliminating the need for explicit BVDM knowledge, which is inherently inferred from UE data reports and raytracing simulations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional CSI reporting methods are used to accurately report multiple channel components, then channel state information accuracy is improved, but uplink overhead and resource consumption increase significantly
Solution Approach 1:
The patent extracts only the essential channel information that cannot be predicted, by separating predictable channel components (modeled through prior channel information and location data) from unpredictable variations (reported as residual information). This extraction approach transmits only the necessary deviation from the predicted channel state.
Solution Approach 2:
The system performs preliminary channel prediction using prior channel information and location data before actual CSI reporting. The gNB predicts channel characteristics in advance based on historical data and UE location, then only reports deviations from this prediction, significantly reducing reporting overhead.
2Measurement precision
If traditional CSI reporting methods are used to accurately report multiple channel components, then channel state information accuracy is improved, but uplink power consumption increases
Solution Approach 1:
The patent extracts only the essential channel information that cannot be predicted, by separating predictable channel components (modeled through prior channel information and location data) from unpredictable variations (reported as residual information). This extraction approach transmits only the necessary deviation from the predicted channel state.
Solution Approach 2:
The system performs preliminary channel prediction using prior channel information and location data before actual CSI reporting. The gNB predicts channel characteristics in advance based on historical data and UE location, then only reports deviations from this prediction, significantly reducing reporting overhead.
3Quantity of substance
If VAE-based information bottleneck method is used to reduce reporting overhead, then uplink overhead is reduced, but channel state information accuracy deteriorates
Solution Approach 1:
The system performs preliminary channel prediction using prior channel information and location data before actual CSI reporting. The gNB predicts channel characteristics in advance based on historical data and UE location, then only reports deviations from this prediction, significantly reducing reporting overhead.
Solution Approach 2:
The patent changes the reporting parameter from full channel state information to residual channel information (difference between actual and predicted channel). This parameter transformation maintains accuracy while reducing overhead, as the residual contains only the unpredictable variations.
4Measurement precision
If explicit BVDM knowledge is used for channel prediction, then prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The system uses the UE's own location information and prior channel measurements to enable channel prediction, rather than requiring external BVDM data. The UE and gNB leverage their existing data to predict channel characteristics, eliminating the need for separate digital twin infrastructure.
Data Source
AI summary
Method, comprising receiving a terminal location information or a location-like information from a terminal; selecting one or more first pairs of prior channel information among one or more stored first pairs of prior channel information based on the terminal location information or the location-like information, respectively; inputting the terminal location information or the location-like information, respectively, and the selected one or more first pairs of prior channel information into a trained interpolation neural network to obtain a first estimation of a channel between the terminal and a base station as an output from the interpolation neural network; providing the weights of the trained neural network to the terminal; wherein each of the one or more first pairs of prior channel information comprises a location information related to a respective prior channel and a first representation of the respective prior channel.