Zero-Forcing Precoding Optimization via Dynamic Sectorization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Massive MIMO systems face significant computational burdens due to the cubic complexity of matrix inversion in Zero-Forcing Precoding, making it unfeasible for large numbers of user terminals, especially when reassessments are required frequently.
Innovation Solution
Implementing dynamic sectorization and higher directivity antenna patterns to reduce computational complexity by transforming dense channel matrices into sparse representations, using backplane sectorization and Minimum Degree Algorithm for sparse Cholesky decomposition, and adjusting antenna directivity to create more zeros in the channel matrix.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If Zero-Forcing precoding is implemented in Massive MIMO systems, then spectral efficiency is improved, but computational complexity increases cubically with the number of users
Solution Approach 1:
The patent segments the channel matrix into sparse and dense components, applying different processing strategies to each. The sparse representation captures the dominant signal paths while ignoring negligible components, thereby reducing the dimensionality of the matrix inversion problem and its associated cubic complexity.
Solution Approach 2:
The patent transforms the channel matrix from its original dense form into a sparse representation by modifying its structural parameters. This sparsification changes the computational characteristics of subsequent operations, enabling more efficient precoder calculations while preserving the essential signal characteristics needed for high spectral efficiency.
2Reliability
If frequent reassessments of channel conditions are performed, then system performance is maintained, but computational burden becomes unfeasible for large numbers of user terminals
Solution Approach 1:
The patent performs preliminary sparsification of the channel matrix before reassessment operations are needed. By pre-processing the channel information into a sparse format, subsequent performance reassessments can be conducted more efficiently, enabling frequent updates without prohibitive computational costs.
Solution Approach 2:
The patent creates a simplified sparse copy of the channel matrix that retains the essential characteristics needed for performance assessment. This compressed representation allows for rapid recalculation of precoders during frequent reassessments, maintaining system reliability while reducing the power and computational resources required.
3Measurement precision
If dense channel matrices are used for accurate signal processing, then precision is maintained, but computational efficiency decreases
Solution Approach 1:
The patent applies local quality by treating different components of the channel matrix differently. Rather than uniformly processing all elements, it identifies and retains only the significant local components (sparse elements) while discarding or approximating the negligible ones. This selective processing maintains precision where it matters most while improving overall computational efficiency.
Solution Approach 2:
The patent changes the structural parameters of the channel matrix from dense to sparse format. This transformation modifies the density parameter while preserving the signal-to-noise ratio and other critical signal characteristics, thereby maintaining processing precision while achieving the computational efficiency needed for large-scale Massive MIMO systems.
Data Source
AI summary
Massive MIMO systems provide impressive spectral efficiencies through beam forming techniques such as Zero-Forcing Precoding (ZFP). Unfortunately, ZFP imposes a considerable computational burden for each additional user. Relationships between the antennas, the users, and the environment must be rapidly, and accurately, reassessed during ZFP on an ongoing basis. Brute force approaches to these reassessments may be unfeasible for certain hardware and design conditions. Accordingly, various of the proposed embodiments implement representational optimizations which reduce the computational burden for each reassessment. Some embodiments employ “dynamic sectorization”, whereby the serviced environment is divided into regions and the corresponding representation is modified to reduce the computations of each reassessment. A backplane, antenna separation/directivity and thresholds for environment noise may each be adjusted to reduce the computational burden.