5G NR Beam Pair Selection Using Federated Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-frequency wireless communication in 5G NR networks faces challenges with low reliability due to severe propagation loss, necessitating the formation of highly directional signals, but finding optimal signal pairs is hindered by high latency as devices constantly change location.
Innovation Solution
Utilizing neural networks trained through federated learning and adaptive pruning to identify optimal beam pairs by aggregating weight information from multiple edge devices, minimizing latency and improving accuracy in beam pair selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If devices form highly directional signals to improve reliability in high-frequency communication, then reliability is improved, but latency increases when finding optimal signal pairs
Solution Approach 1:
The system pre-trains neural networks at multiple edge devices using federated learning to predict optimal beam pairs before actual communication needs arise. When location changes occur, the pre-trained models can quickly identify new optimal signal pairs without extensive real-time computation, thus reducing latency while maintaining reliability
Solution Approach 2:
Multiple edge devices create local copies of neural network models through federated learning. Each device has its own copy that can independently predict optimal beam pairs, eliminating the need for centralized computation and reducing latency when devices move or conditions change
2Adaptability or versatility
If devices constantly update optimal signal pair selection to adapt to location changes, then adaptability is improved, but computational complexity and latency increase
Solution Approach 1:
The patent replaces traditional mechanical/computational beam pairing methods with neural network-based prediction. Instead of exhaustively searching through possible beam pairs using complex algorithms, the system uses trained neural networks that automatically predict optimal pairs, significantly reducing computational complexity while maintaining adaptability to location changes
Solution Approach 2:
Each edge device independently maintains and updates its own neural network model through federated learning, performing computations locally rather than relying on centralized processing. This self-service approach reduces overall system complexity and enables faster local adaptation to changing conditions
Data Source
AI summary
Apparatuses, systems, and techniques to identify a first one or more fifth generation new radio (5G NR) signal pairs having one or more quality characteristics based, at least in part, on neural network weight information received from a plurality of different sources. In at least one embodiment, neural network weight information can be aggregated to generate one or more neural networks to identify the first one or more 5G NR signal pairs.