Kalman Filter Signal Aggregation in Distributed MIMO Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for combining and aggregating signals in distributed Multiple Input Multiple Output (D-MIMO) networks are unscalable, computationally complex, and have low performance due to high fronthaul capacity requirements and sequential processing limitations.
Innovation Solution
The implementation of Kalman filters in a D-MIMO network to aggregate signals from different access points, using two types of Kalman filters for updating signal estimates and combining covariance matrices, which enables scalable and efficient signal processing across various network topologies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing methods for combining and aggregating signals in D-MIMO networks are used, then signal aggregation is achieved, but computational complexity increases and performance decreases
Solution Approach 1:
The patent transforms the signal aggregation problem by changing the mathematical approach from traditional matrix inversion to Kalman filter-based recursive estimation. This parameter change in the processing methodology reduces computational complexity from O(L^3) to O(L^2) while maintaining aggregation performance across different network topologies
Solution Approach 2:
The patent segments the signal aggregation process into two distinct types: Type 1 Kalman filter for basic signal estimation and Type 2 Kalman filter for covariance matrix combination. This segmentation allows independent optimization of each processing stage, improving overall efficiency while reducing the computational burden on any single component
2Productivity
If existing methods for combining and aggregating signals in D-MIMO networks are used, then signal aggregation is achieved, but fronthaul capacity requirements increase
Solution Approach 1:
The patent extracts only the essential information needed for signal aggregation by using Kalman filters to process and compress data at each access point before transmission over the fronthaul. This extraction approach sends condensed state estimates and covariance matrices rather than raw signal data, significantly reducing fronthaul capacity requirements while preserving aggregation performance
3Productivity
If existing methods for combining and aggregating signals in D-MIMO networks are used, then signal processing is performed, but processing limitations due to sequential processing increase
Solution Approach 1:
The patent creates a universal signal aggregation framework using Kalman filters that functions across multiple network topologies (star, mesh, hierarchical) without requiring topology-specific processing logic. The same Type 1 and Type 2 Kalman filter combination works universally, providing processing flexibility and ease of operation across different deployment scenarios
Data Source
AI summary
A method by a first network node for handling data. The first network node operates in a communications network. The first network node applies a Kalman filter of a first type to a first aggregation of a first set of measurements (yl) collected via a first plurality of antenna elements, and a first set of filtered data (ŝ0) received from a second network node, using a first covariance matrix (P0). The applying of the Kalman filter () outputs a second set of filtered data (ŝ), and a second covariance matrix (P). The first network node then sends, in uplink, the second set of filtered data (ŝ) and the second covariance matrix (P) to one of: another network node subsequently adjacent to the first network node towards a central network node, and the central network node.