CRAM Projection Matrix Subspace Extraction for Low PAPR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing Convex Reduction of Amplitude (CRAM) framework for low peak-to-average power ratio (PAPR) precoding in Massive MIMO systems faces challenges with high throughput requirements for the CRAM coefficient interface, leading to increased complexity and power consumption due to the need to transfer large projection matrices.
Innovation Solution
The use of a signal subspace to generate projection matrices reduces the throughput requirement by transmitting only the signal subspace, eliminating the need for the CRAM coefficient interface and allowing for local computation of projection coefficients, thereby reducing the computational and power demands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional CRAM projection matrices are transferred over the CRAM coefficient interface, then low PAPR precoding is achieved, but the throughput requirement becomes excessively high (449.28 Gbps) and complexity increases
Solution Approach 1:
The patent extracts and transmits only the essential signal subspace information (eigenvectors) rather than the complete CRAM projection matrices. By identifying and transmitting only the K dominant eigenvectors of the channel matrix, the system achieves the same low PAPR precoding performance while dramatically reducing the data transmission requirement from 449.28 Gbps to a much lower rate, thereby reducing interface complexity
Solution Approach 2:
The patent changes the parameter being transmitted from full projection matrices to reduced-dimensional subspace representations. By transforming the complete matrix information into compact eigenvector forms and utilizing the rank-K property of MIMO channels, the system maintains functional equivalence while reducing the throughput requirement by an order of magnitude
2Measurement precision
If full CRAM projection matrices are transferred, then accurate precoding is achieved, but power consumption increases due to high throughput requirements
Solution Approach 1:
The patent extracts only the critical subspace information (signal and interference eigenvectors) needed for accurate precoding, eliminating the transmission of redundant matrix elements. This extraction approach maintains precoding accuracy by preserving the essential channel characteristics while reducing power consumption associated with high-speed data transmission
Solution Approach 2:
The patent segments the channel information into distinct components: signal subspace eigenvectors and interference subspace eigenvectors. By separately identifying and transmitting these segmented components, the system achieves accurate precoding through coordinated use of both subspaces while minimizing the total throughput requirement and associated power consumption
3Loss of information
If large projection matrices are transmitted, then complete channel information is provided, but the number of SERDES lanes and system complexity increase
Solution Approach 1:
The patent extracts the essential channel information in the form of signal subspace eigenvectors and interference subspace eigenvectors, discarding redundant elements. This extraction maintains channel information completeness for precoding purposes while reducing the transmitted data volume, thereby decreasing the number of SERDES lanes required and simplifying the overall system architecture
Solution Approach 2:
The patent transforms the channel information from a high-dimensional full matrix representation to a lower-dimensional subspace representation. By utilizing the inherent rank-K structure of MIMO channels and representing the channel through its dominant eigenvectors, the system achieves dimensionality reduction that preserves essential information while reducing the throughput requirement and SERDES lane complexity
Data Source
AI summary
A network node is provided. The network node includes processing circuitry configured to determine at least one Convex Reduction of Amplitude, CRAM, projection matrix based at least in part on a signal subspace of scheduled wireless devices, and optionally cause transmission based at least in part on the at least one CRAM projection matrix.