Antenna Panel Switching Using Long-Term Reward Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In wireless communication systems, particularly in 5G networks, terminal devices face challenges in optimizing antenna panel usage due to limited resources, leading to suboptimal signal quality and increased hardware complexity, which affects user experience and network performance.
Innovation Solution
Implementing a machine learning algorithm to select the most suitable antenna panel from a plurality based on received signals, determining a long-term reward value, and adjusting antenna panel usage dynamically to maximize signal quality and minimize radio link failures and handover failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple antenna panels are used to improve signal quality and reliability, then communication performance is improved, but device complexity and resource consumption increase
Solution Approach 1:
The terminal device autonomously performs antenna panel selection using machine learning algorithms without requiring complex network control or manual configuration. The device self-manages the selection process by evaluating long-term reward values and making independent decisions, thereby reducing overall system complexity while maintaining multiple antenna panels for improved reliability
Solution Approach 2:
The patent transforms the antenna panel selection problem into a parameter optimization problem by using machine learning to evaluate and compare long-term reward values associated with different antenna panels. This allows the system to dynamically adjust selection based on learned parameters rather than static configurations, improving signal quality while managing complexity through intelligent parameter management
2Measurement precision
If machine learning algorithms are implemented for antenna panel selection, then selection accuracy and long-term performance are improved, but computational resources and processing time are consumed
Solution Approach 1:
The machine learning algorithm performs partial computation by calculating long-term reward values only for relevant antenna panels based on current communication conditions. Rather than exhaustively evaluating all possible panels continuously, the system performs selective computations when needed, reducing energy consumption while maintaining selection accuracy through targeted machine learning operations
Solution Approach 2:
The system maintains continuous learning and selection capability by accumulating long-term reward data over time. The machine learning model continuously updates its understanding of antenna panel performance, allowing it to make accurate selections while optimizing energy usage patterns through sustained, efficient computation rather than intermittent heavy processing
3Productivity
If dynamic antenna panel switching is performed to optimize throughput and latency, then network performance is improved, but radio link failures and handover failures may increase
Solution Approach 1:
The machine learning algorithm incorporates feedback mechanisms by continuously evaluating long-term reward values that reflect both throughput performance and link stability. The system learns from past switching decisions and their outcomes, adjusting future selections to balance throughput optimization with maintaining radio link stability, thereby reducing failures while preserving performance gains
Solution Approach 2:
The system performs preliminary evaluation of antenna panels by calculating and storing long-term reward values before actual switching decisions are made. This advance preparation allows the system to select panels that have been pre-assessed for both performance and stability, reducing the risk of failures during dynamic switching while maintaining optimized throughput through informed selection
Data Source
AI summary
Disclosed is a method comprising using a machine learning algorithm to select an antenna panel from a plurality of antenna panels. A first long-term reward value associated with the selected antenna panel is determined based at least partly on one or more first signals received on the selected antenna panel. A second signal is then transmitted or received via the selected antenna panel, if the first long-term reward value exceeds one or more second long-term reward values associated with at least a subset of the plurality of antenna panels.