Kalman Filter Adaptive Beamforming Computation Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Adaptive beamforming algorithms require significant computational resources due to the need for large numbers of computing operations and iterations, leading to high power consumption and costs in 5G MU-MIMO applications, especially in estimating and updating antenna array beam weights.
Innovation Solution
The method employs a Kalman filter algorithm to estimate and update a linear model of unwrapped beam weights, reducing the computational resources required by 80 to 90% by using a reduced set of data, typically 5-20% of the initial samples, and applying a state estimation filter to calculate new beam pattern weights efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If adaptive beamforming algorithms are used to compute optimal beam patterns, then signal transmission quality is improved, but computation resource consumption increases significantly
Solution Approach 1:
The patent segments the beam pattern computation into two parts: (1) initial beam pattern computation using full data set, and (2) incremental updates using only new received signal samples. This segmentation allows the system to benefit from adaptive beamforming quality while reducing computational complexity by avoiding redundant processing of already-analyzed samples.
Solution Approach 2:
The patent performs preliminary computation of the initial beam pattern weights using the full data set before incremental updates. This preliminary action establishes a baseline that eliminates the need to reprocess historical data during subsequent updates, significantly reducing ongoing computational requirements while maintaining transmission quality.
2Measurement precision
If adaptive beamforming algorithms iterate multiple times for convergence, then beam pattern accuracy is improved, but computation time increases
Solution Approach 1:
The patent implements periodic computation where full beam pattern re-computation occurs only when necessary (periodically), while incremental updates are performed continuously. This periodic action maintains beam pattern accuracy through regular full computations while minimizing computation time between these periodic events by using efficient incremental updates.
Solution Approach 2:
The patent uses feedback from the initial beam pattern computation to guide incremental updates. The initial computation provides reference weights that serve as feedback for subsequent updates, allowing the system to maintain accuracy with fewer iterations by building upon previously established accurate estimates rather than starting from scratch each time.
3Measurement precision
If full data sets are used for beam weight estimation, then estimation accuracy is improved, but data processing load increases
Solution Approach 1:
The patent extracts and processes only the incremental portion of new received signal samples that have arrived since the last computation, rather than reprocessing the entire data set. This extraction approach maintains estimation accuracy by incorporating all necessary new information while eliminating redundant processing of historical data, significantly reducing data processing load.
Solution Approach 2:
The patent applies partial action by processing only the necessary subset of data (new samples since last update) rather than the complete data set. This partial processing is sufficient to maintain accurate beam weight estimates because it focuses computational effort on the incremental changes that actually affect the current beam pattern requirements.
Data Source
AI summary
A method for reducing adaptive beam forming computation resources for estimating and updating a model of unwrapped beam weights. The optimal beam pattern weights of an antenna array are estimated using an adaptive beamforming algorithm. An initial model is created for either magnitude or phase components of the optimal beam pattern weights computed from the adaptive beamforming algorithm estimates. For each time step, a measurement of optimal beam pattern weights is estimated, using a reduced set of data comprising 5-20% of first samples of signal reference data. New beam pattern weights are computed using a magnitude Kalman filter (KF) and/or phase KF, wherein the computation resources required to obtain the new beam pattern weights are reduced by 80 to 90% over an adaptive beam forming algorithm.