Base Station ANN Configuration for Multi-Terminal Wireless Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile communication systems lack efficient methods for configuring and training artificial neural networks (ANNs) when multiple terminals coexist within a cell, limiting the effectiveness of AI/ML applications for channel state information feedback, beam management, and positioning accuracy.
Innovation Solution
The proposed solution involves defining different types of ANN sharing and learning configurations, including units of cells, terminal groups, terminals, and cell groups, with mechanisms for capability reporting, online learning, and weight vector updates between base stations and terminals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If artificial neural networks are configured for each terminal individually, then learning accuracy and positioning precision are improved, but device complexity and management difficulty increase significantly
Solution Approach 1:
The patent segments ANN configuration management into multiple levels: cell-level common configuration, terminal-group-specific configuration, and individual terminal configuration. This hierarchical segmentation allows positioning precision to be maintained through terminal-specific parameters while reducing management complexity through automated group-based templates and base station centralized control.
Solution Approach 2:
The base station acts as an intermediary that automatically generates, distributes, and manages ANN configuration information across multiple terminals. It receives capability information from terminals, determines appropriate configuration parameters, and transmits configuration information to terminals, thereby reducing direct management complexity while maintaining individual terminal optimization.
2Productivity
If ANN configuration information is transmitted to each terminal, then learning capability and CSI feedback effectiveness are improved, but communication overhead and signaling load increase
Solution Approach 1:
The patent merges common ANN configuration parameters at the cell level and applies them to multiple terminals simultaneously through terminal groups. This reduces communication overhead by transmitting a single set of common configuration information to serve multiple terminals, while still allowing individual terminal adjustments when needed.
Solution Approach 2:
The patent applies local quality by transmitting detailed ANN configuration information only to terminals that require it, while other terminals use default or group-level configurations. Capability information reporting allows the base station to identify which terminals need individualized configuration, reducing overall communication overhead while maintaining learning capability where necessary.
3Adaptability or versatility
If capability information reporting is implemented for ANN support, then system adaptability and configuration accuracy are improved, but signaling overhead and processing time increase
Solution Approach 1:
The patent implements capability information reporting as a preliminary action during initial terminal registration or before ANN configuration is needed. Terminals report their ANN support capabilities in advance, allowing the base station to prepare appropriate configuration information without delay when actual ANN operations are required, thus minimizing processing time during critical operations.
Data Source
AI summary
A method of a base station may comprise: receiving, from each of terminals belonging to the base station, a capability information report message related to an artificial neural network; generating artificial neural network configuration information based on capability of the base station and the capability information report message received from each of the terminals; and transmitting the generated artificial neural network configuration information to the terminals belonging to the base station.