Reciprocity-Aided Beamspace Precoding with Iterative Inversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The computational complexity of interference-aware downlink precoder calculation in massive MIMO systems with large numbers of antennas is high due to the need for inversion of the interference covariance matrix, which is impractical for active antenna systems.
Innovation Solution
Iterative estimation of the inverse of the interference covariance matrix combined with beamspace dimension reduction to reduce computational complexity, avoiding direct inversion of the full-dimension matrix.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If interference-aware downlink precoder calculation is implemented in massive MIMO systems, then inter-cell interference suppression performance is improved, but computational complexity increases due to matrix inversion requirements
Solution Approach 1:
The patent segments the full-dimension interference covariance matrix into multiple block matrices based on antenna groups. Instead of inverting one large M×M matrix, the system divides antennas into groups and inverts smaller block matrices separately, significantly reducing computational complexity while maintaining interference suppression performance.
Solution Approach 2:
The patent transforms the problem from the spatial domain to the beam domain using beamspace transformation. By rotating the coordinate system and working in the beam domain, the system exploits the low-rank structure of the interference covariance matrix, enabling efficient inversion through eigenvalue decomposition and reducing computational burden.
2Measurement precision
If full-dimension interference covariance matrix inversion is performed, then precoding accuracy is improved, but implementation becomes impractical for active antenna systems with large numbers of antennas
Solution Approach 1:
The patent divides the antenna array into multiple groups and segments the interference covariance matrix into block matrices. Each block corresponds to a specific antenna group, allowing separate inversion operations that are computationally feasible while preserving the overall precoding accuracy through proper reconstruction of the full precoder.
Solution Approach 2:
The patent introduces beamspace transformation as an intermediary step between channel estimation and precoder calculation. This transformation diagonalizes the problem structure and reveals the low-rank nature of interference, serving as a mathematical mediator that enables accurate yet computationally efficient inversion through eigenvalue decomposition.
Data Source
AI summary
According to some embodiments, a method performed by a network node for interference aware downlink transmission includes determining a reciprocity-aided interference aware transmission precoder based on a downlink wideband interference covariance matrix estimated from a plurality of received sounding reference signals and based on a matrix inversion associated with the downlink wideband interference covariance matrix. The matrix inversion is determined based on an iterative inverse covariance estimation. The method further includes transmitting a downlink signal using the reciprocity-aided interference aware transmission precoder.