Dynamic Spectrum Access via Reinforcement Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing dynamic spectrum access (DSA) solutions rely on predetermined communication channels and require high computational power, making them inefficient and unsuitable for real-time adaptation in fluctuating environments, especially in safety-critical systems where human-readable data transparency is necessary.
Innovation Solution
A method using deep reinforcement learning and machine learning-based spectrum analysis to identify optimal carrier frequencies and bandwidths for transmitting information, allowing for real-time dynamic spectrum allocation and human-readable data interpretation, leveraging programmable integrated photonics for microwave spectrum characterization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If deep reinforcement learning and machine learning-based spectrum analysis are used to identify optimal carrier frequencies and bandwidths, then dynamic spectrum allocation efficiency is improved, but computational power requirements increase
Solution Approach 1:
The system performs preliminary spectrum analysis by identifying carrier frequencies and bandwidths of existing transmissions before allocating new spectrum spaces. This advance preparation allows the reinforcement learning model to make informed allocation decisions without requiring excessive computational power during real-time transmission, resolving the contradiction between allocation efficiency and computational requirements
Solution Approach 2:
The patent replaces traditional mechanical spectrum allocation methods with machine learning-based analysis and reinforcement learning for optimization. This substitution enables more efficient spectrum utilization through intelligent pattern recognition while the system manages computational resources through selective analysis of spectrum characteristics rather than exhaustive processing
2Extent of automation
If fully automated DSA solutions using RNN are implemented, then access decision speed is improved, but data transparency and human readability are lost
Solution Approach 1:
The reinforcement learning implementation includes feedback mechanisms that provide human-readable information about spectrum allocation decisions. The system maintains transparency by feeding back interpreted spectrum characteristics and allocation rationale in formats that humans can understand, while still achieving automated decision-making. This resolves the contradiction by allowing full automation while preserving information transparency through structured feedback outputs
3Device complexity
If predetermined communication channels are used for DSA, then system complexity is reduced, but spectrum utilization efficiency decreases
Solution Approach 1:
The patent implements dynamic spectrum allocation where the system adapts to changing spectrum conditions by identifying optimal carrier frequencies and bandwidths in real-time. Rather than using fixed predetermined channels, the system dynamically adjusts allocation based on current spectrum usage patterns detected through machine learning analysis. This dynamic approach improves spectrum utilization efficiency while maintaining manageable system complexity through automated adaptation
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of allocating space on a spectrum on which information is transmitted, to information to be transmitted, the method comprising: identifying carrier frequencies and bandwidths of information being transmitted on the spectrum, determining an optimal carrier frequency and bandwidth for the information to be transmitted based on the carrier frequencies and bandwidths of information being transmitted on the spectrum; transmitting the information to be transmitted using modulation at the identified optimal carrier frequency and bandwidth.