Cell Deactivation Control Using Predicted Throughput Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems face challenges in optimizing energy consumption while maintaining communication quality, particularly in distributed self-organized networks where deactivating cells can lead to increased traffic on remaining frequency layers, resulting in reduced throughput below threshold levels.
Innovation Solution
Implementing a service management and orchestration framework that utilizes machine learning models to predict key performance indicators, allowing for the activation or deactivation of cells based on predicted throughput changes, with configurable confidence levels to maintain throughput above thresholds and reduce power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If cells are deactivated to reduce energy consumption, then power consumption decreases, but throughput may fall below threshold levels
Solution Approach 1:
The system performs preliminary actions by predicting future throughput values before making cell activation/deactivation decisions. The machine learning model forecasts throughput for multiple future time instances, allowing the system to proactively identify when deactivation will maintain throughput above thresholds, thus preventing performance degradation before it occurs.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring observed throughput values and comparing them against predicted values. This feedback loop allows the system to validate predictions, adjust to actual network conditions, and make informed decisions about cell activation and deactivation that balance energy consumption with throughput requirements.
2Loss of energy
If machine learning models predict throughput changes to guide cell activation/deactivation, then energy consumption is optimized, but system complexity increases
Solution Approach 1:
The machine learning model serves as an intermediary between raw network observations and cell activation/deactivation decisions. It transforms complex observed performance indicators into simplified predicted throughput values, enabling automated decision-making without requiring complex manual analysis or control logic.
Solution Approach 2:
The system implements self-service by using automated machine learning-based predictions to make cell activation and deactivation decisions without human intervention. The service management function automatically processes observed indicators, generates predictions, and determines activation states, reducing operational complexity and enabling energy optimization through autonomous control.
Data Source
AI summary
This disclosure provides systems, methods, and apparatus for energy saving using predicted performance indicators. The described techniques may enable a service management and orchestration framework (SMO) to determine one or more cells to deactivate based on one or more predicted key performance indicators (KPIs) associated with deactivating the cells. For example, the SMO may use a machine learning (ML) model, a neural network (NN), and the like to predict how a throughput associated with the one or more cells may change based on deactivating a first cell for a first period of time. In some aspects, the SMO may therefore predict a first time period or a threshold load for to deactivate the first cell that may result in a throughput of one or more other cells decreasing an amount that is less than a threshold throughput decrease.