LSAS Base Station Calibration Fat-Tree Architecture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale antenna systems (LSAS) face impracticalities in calibration due to the need for a reference antenna, which becomes inefficient as the number of service antennas increases, leading to prolonged calibration times and placement challenges.
Innovation Solution
A fat-tree architecture is implemented, where each node in the system calibrates with its children, starting from the central controller, eliminating the need for a reference antenna and reducing calibration time by allowing simultaneous calibration of neighbor ratios across levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a reference antenna is used for calibration in conventional techniques, then calibration can be performed, but calibration time increases proportionally with the number of service antennas and placement becomes impractical for large-scale systems
Solution Approach 1:
The patent segments the calibration process by introducing intermediate relay antennas that divide the calibration path into smaller segments. Instead of direct calibration between the reference antenna and each service antenna, the system uses relay antennas to mediate the calibration signals, enabling parallel calibration operations and reducing overall calibration time while maintaining accuracy.
Solution Approach 2:
The patent introduces relay antennas as intermediary elements in the calibration process. These relay antennas receive calibration signals from the reference antenna and forward them to service antennas, enabling indirect calibration paths that allow simultaneous calibration of multiple antennas and reduce the sequential time required.
2Measurement precision
If a reference antenna is used for calibration, then calibration can be performed, but the system cannot serve terminals during the calibration process
Solution Approach 1:
The calibration process is segmented into multiple independent calibration paths using relay antennas. This segmentation allows different parts of the antenna array to be calibrated simultaneously while others continue serving terminals, reducing the impact on overall system productivity and terminal service availability.
Solution Approach 2:
The patent implements partial calibration by using relay antennas to calibrate subsets of service antennas in parallel. This allows the calibration process to proceed without requiring complete system shutdown, as calibration can be performed on portions of the array while other portions remain operational for terminal service.
3Reliability
If the number of service antennas is increased to improve network performance, then beam sharpness and transmission discrimination improve, but calibration becomes increasingly impractical
Solution Approach 1:
The patent segments the large antenna array into multiple calibration groups that can be calibrated in parallel using relay antennas. This segmentation makes the calibration process scalable to large numbers of service antennas by dividing the complex calibration task into smaller, manageable, and simultaneous operations.
Solution Approach 2:
The patent introduces an additional spatial dimension to the calibration process by deploying relay antennas at intermediate positions. This creates a hierarchical calibration structure with multiple levels (reference antenna → relay antennas → service antennas), enabling parallel calibration operations that scale with the number of service antennas.
Data Source
AI summary
In one embodiment, an LSAS base station has a tree architecture including a central controller at the tree root, one or more intermediate levels of hubs forming tree branches, and an antenna module for each service antenna at a tree leaf. To calibrate an overall ratio characterizing differences between downlink and uplink channels between the central controller and each antenna module, neighbor-calibration procedures are calibrate neighbor ratios between parent nodes and their child nodes. Appropriate neighbor ratios are multiplied to generate overall ratios. To reduce time required to perform the neighbor-calibration procedures, odd and even phases are performed. In the odd phase, neighbor ratios are generated between parents in all odd levels and their children in even levels, and analogously for the even phase for even-level parents and odd-level children. Within each phase, independent parent-child pairs can be calibrated simultaneously.