Cellular SON Configuration Using Signal Deviation Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cellular communication systems face performance degradation due to unreliable forecast models caused by inconsistent historical KPI parameters resulting from network configuration changes, leading to inaccurate predictions and difficulties in implementing Self-Organizing Network (SON) functionalities, especially in 5G systems where users transmit on different frequency bands.
Innovation Solution
A method that combines KPI parameters with geo-localized tracing/reporting data to optimize network configuration by determining estimated performance indicators, using artificial neural networks and optimization algorithms to adjust cell parameters, and iteratively optimize network configurations based on signal level deviations and user distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If forecast procedures use historical KPI parameters collected over long monitoring time periods, then forecast accuracy improves, but network configuration changes cause inconsistent data that reduces reliability
Solution Approach 1:
The system performs preliminary actions by collecting and storing KPI parameters and configuration data before forecast issues arise. It proactively identifies network configuration changes and adjusts the forecast model accordingly, preventing the inconsistency problem rather than reacting to it after it occurs.
Solution Approach 2:
The system changes parameters by dynamically adjusting the forecast model based on detected network configuration changes. When configuration changes are identified, the system modifies how historical data is weighted or processed, ensuring that the forecast remains accurate despite changes in network topology or parameters.
2Reliability
If KPI parameters are monitored and processed to detect performance issues, then network performance optimization is achieved, but the long processing time causes performance degradation
Solution Approach 1:
The system performs preliminary analysis by continuously collecting and pre-processing KPI parameters and configuration data in the background. When performance issues are detected, the forecast model can immediately provide predictions because the data preparation work was already done in advance, reducing the effective response time.
Solution Approach 2:
The system implements feedback mechanisms where forecast results and actual performance data are continuously compared. This feedback loop allows the system to learn from past performance and adjust its monitoring and processing strategies, optimizing the balance between detection speed and accuracy over time.
3Productivity
If network configuration changes are made to optimize performance, then user distribution improves, but forecast model reliability deteriorates due to data inconsistency
Solution Approach 1:
The system handles parameter changes by detecting network configuration changes and adapting the forecast model accordingly. When configuration parameters change, the system adjusts the data processing methodology to account for these changes, ensuring that the forecast model remains reliable despite the underlying network changes that are necessary for optimization.
Solution Approach 2:
The system applies dynamics by making the forecast model adaptive rather than static. The model dynamically adjusts its parameters and data processing methods based on current network configuration states, allowing it to maintain accuracy through periods of network optimization and configuration changes.
Data Source
AI summary
A method for configuring a cellular network comprising a plurality of network cells. The method comprises:monitoring, during a monitoring time period, a performance indicator of each network cell for each one of a plurality of monitoring time intervals of said monitoring time period;determining a plurality of sets of measured signal levels, wherein each set of measured signal levels is associated with a respective territorial portion of a geographic area covered by the cellular network, and wherein each measured signal level of each set of measured signal levels is associated with a respective measured network cell, among the plurality of network cells, at least partially within that territorial portion, with a respective cell configuration of said measured network cell, and with a respective monitoring time interval;determining, for each measured network cell, a simulated signal level being indicative of an expected signal level for that measured network cell in the respective cell configuration;determining, for each measured network cell and for each monitoring time interval, a signal level deviation as a difference between the respective measured signal level and the respective simulated signal level;for a reference network configuration (RNC1) including, for each measured network cell, a respective reference cell configuration taken by that measured network cell after the monitoring time period:(i) determining, for each measured network cell, a reference simulated signal level indicative of an expected signal level for that measured network cell in the reference cell configuration;(ii) updating, for each monitoring time interval, each measured signal level according to the respective simulated signal level and to the respective signal level deviation;(iii) determining, for each measured network cell and for each monitoring time interval, an estimated performance indicator based on the respective performance indicator, on a number of sets of measured signal levels having at least one measured signal level associated with said measured network cell, and on the respective updated measured signal level, andconfiguring the cellular network according to the estimated performance indicators determined for the measured network cells.