AI/ML Model Activation in Wireless Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Next-generation wireless access networks (NR) require flexible frame structures and efficient radio resource multiplexing to meet diverse QoS requirements for different usage scenarios, such as eMBB, mMTC, and URLLC, which existing technologies have not adequately addressed.
Innovation Solution
The implementation of an AI/ML model in wireless communication networks, allowing for activation or deactivation based on instruction information, and utilizing a transmitter, receiver, and controller to manage AI/ML operations, enabling efficient use of AI/ML in NR systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If AI/ML model is activated to improve resource management and performance optimization, then network adaptability and productivity are improved, but device complexity and energy consumption increase
Solution Approach 1:
The patent introduces a network controller as an intermediary that manages and coordinates AI/ML model operations across the network. This centralizes the complexity of model activation, parameter tuning, and resource allocation, allowing individual network elements to benefit from AI/ML capabilities without each bearing the full complexity burden. The controller mediates between the need for intelligent optimization and the practical constraints of device complexity.
Solution Approach 2:
The patent segments the AI/ML functionality into modular components that can be selectively activated based on network conditions and requirements. Rather than deploying monolithic AI/ML systems throughout the network, the approach divides intelligence into discrete, manageable units that can be independently controlled and optimized, reducing overall system complexity while maintaining adaptability.
2Productivity
If AI/ML model is activated to optimize resource management, then productivity is improved, but energy consumption increases
Solution Approach 1:
The patent implements periodic activation and deactivation of AI/ML models based on network traffic patterns and resource utilization thresholds. Rather than continuous operation, the models are activated only when network conditions warrant intelligent optimization, and deactivated when standard operations suffice. This periodic approach maintains productivity benefits while significantly reducing energy consumption during low-demand periods.
Solution Approach 2:
The patent dynamically adjusts operational parameters of AI/ML models based on real-time network conditions, including activation thresholds, model complexity levels, and resource allocation parameters. By changing these parameters adaptively, the system optimizes the balance between productivity gains and energy consumption, scaling AI/ML usage to match actual network needs rather than operating at fixed high levels.
Data Source
AI summary
The present embodiments provide a method by which a terminal uses an AI/ML model in a wireless communication network, the method comprising the steps of receiving activation or deactivation instruction information for the AI/ML model; activating or deactivating the AI/ML model on the basis of the instruction information; and transmitting a response to activation or deactivation of the AI/ML model.