Deep RL Spectrum Sharing via Receiver Link Quality Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face inefficiencies in spectrum sharing due to reliance on receiver feedback and lack of consideration for receive link quality, leading to suboptimal transmitter sensing and potential network overhead.
Innovation Solution
A deep neural network reinforcement learning approach is employed to improve transmitter sensing by using receiver link quality information, where an artificial neural network determines transmission decisions and parameters based on a reward model that maximizes serving rate across transmission devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional receiver feedback-based spectrum sharing is used, then transmission decisions can be made, but spectrum sharing efficiency deteriorates due to lack of link quality consideration
Solution Approach 1:
The patent implements a feedback mechanism where receiver link quality information (such as SINR measurements) is sent back to the transmitter. The neural network uses this feedback to adjust transmission decisions and parameters, creating a closed-loop system that continuously optimizes spectrum sharing based on actual channel conditions, thereby resolving the contradiction between efficiency and reliability.
Solution Approach 2:
The system employs a neural network that autonomously makes transmission decisions based on received link quality information without requiring complex centralized coordination. Each transmitter device independently processes the feedback and adjusts its own transmission parameters, enabling self-optimized spectrum sharing that improves both efficiency and link quality.
2Reliability
If receiver feedback is implemented to improve link quality awareness, then transmission decisions can be optimized, but network overhead increases
Solution Approach 1:
The patent implements a selective feedback mechanism where not all receivers send feedback in every transmission opportunity. Instead, feedback is transmitted only when link quality conditions warrant adjustment or when the neural network determines it is necessary for optimization. This partial feedback approach provides sufficient link quality awareness while minimizing network overhead.
3Productivity
If deep reinforcement learning is used to optimize transmission decisions, then spectrum sharing efficiency improves, but device complexity increases
Solution Approach 1:
The neural network is trained offline using reinforcement learning algorithms before deployment in the actual wireless system. During this preliminary training phase, the network learns optimal transmission strategies through simulated environments. Once trained, the pre-trained neural network is deployed to transmission devices, where it can make decisions based on learned patterns without requiring complex real-time computation, thus reducing operational device complexity while maintaining high spectrum sharing efficiency.
Data Source
AI summary
A method of wireless communication performed by a first transmission device includes determining a set of spectrum sharing parameters based on sensing performed during a sensing period of a current time slot in a fixed contention based spectrum sharing system. The first transmission device shares a spectrum with a second transmission device. The method also includes determining, at a first artificial neural network of the first transmission device, a transmission device action, and/or a transmission parameter in response to receiving the set of spectrum sharing parameters. The method further includes transmitting, from the first transmission device, to a first receiving device during a data transmission phase of the current time slot based on the transmission device action and/or the transmission parameter.