AI/ML Radio Resource Management Cell Subset Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current AI/ML-based Radio Resource Management (RRM) algorithms for mobile communication networks require significant data transfer, storage, and computing resources, leading to high energy consumption and environmental impact, while existing solutions do not effectively account for operator preferences or dynamic cell environments.
Innovation Solution
A method is introduced to dynamically identify a subset of cells for data aggregation and set training parameters based on operator preferences, reducing computational complexity and optimizing Quality of Service (QoS) by training AI/ML models using only data from selected cells, thereby reducing data aggregation and energy usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If AI/ML models are trained using data from all cells in the mobile communication network, then the model accuracy and reliability are improved, but the data transfer requirements, storage capacity, and computing resources increase significantly
Solution Approach 1:
The patent segments the network cells into different categories based on their characteristics (e.g., urban, suburban, rural cells) and trains separate AI/ML models for each segment. This allows the system to maintain high model accuracy for specific cell types while reducing the overall data volume required by focusing computational resources only on relevant segments rather than training a single comprehensive model on all cell data.
Solution Approach 2:
The patent applies local quality by training cell-specific or region-specific AI/ML models that are optimized for local network conditions and characteristics. Each model is tailored to the specific requirements of its target cell or region, allowing for high accuracy in local contexts while avoiding the need to process and store data from the entire network.
2Adaptability or versatility
If AI/ML models are trained using data from all cells, then the model can account for varying network conditions, but the energy consumption and environmental impact increase
Solution Approach 1:
By segmenting the network into distinct cell categories and training separate models for each segment, the system maintains adaptability to varying network conditions without requiring energy-intensive processing of all cell data simultaneously. Each segment can be trained and updated independently, reducing the overall computational energy consumption while preserving the ability to adapt to different network environments.
Solution Approach 2:
The patent applies partial action by training models only on the subset of cell data that is necessary for maintaining adequate model performance, rather than processing all available network data. This selective approach allows the system to maintain adaptability to important network conditions while significantly reducing energy consumption by avoiding unnecessary data processing.
3Productivity
If the AI/ML model processes data from all cells, then comprehensive radio resource management can be achieved, but the computational complexity and processing time increase
Solution Approach 1:
The patent segments the comprehensive radio resource management task into separate, manageable computational units, each handling a specific cell segment. This division reduces the computational complexity of individual processing operations while maintaining overall productivity by parallelizing the management of different cell segments across multiple models.
Solution Approach 2:
The system applies partial action by processing only the subset of cell data necessary for effective radio resource management in each segment, rather than attempting to process all network data simultaneously. This approach maintains adequate management coverage while significantly reducing computational complexity and processing time requirements.
Data Source
AI summary
A device may include a memory configured to store an artificial intelligence or machine learning model (AI/ML) configured to provide an output used in radio resource management of a plurality of cells; and a processor configured to: obtain cell-specific parameters of the plurality of cells of a mobile communication network; select a subset of the plurality of cells based on obtained cell-specific parameters; and cause the AI/ML to be trained with radio access network (RAN)-related data of the subset of the plurality of cells.