Self-Organizing Network Predictive Modeling for Resource Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mobile networks fail to fully utilize detailed user knowledge and specific information to optimize network resource allocation, leading to inefficient network plans and difficulty in predicting outcomes of network reconfiguration without both uplink and downlink information, which hampers performance and interference mitigation.
Innovation Solution
A self-organizing network (SON) system that receives measurement and call trace information from base stations to train predictive models, predicting network performance, estimating missing data, identifying interference, and determining actions to improve network configuration, thereby optimizing resource allocation and reducing interference.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a one-size-fits-all approach with macro cells is used to cover large geographical areas, then network coverage is improved, but network resource allocation efficiency deteriorates
Solution Approach 1:
The network is segmented into multiple types of cells (macro cells, micro cells, pico cells, femto cells) that can be deployed hierarchically. Each cell type serves specific coverage requirements, allowing the network to divide and conquer large geographical areas while maintaining efficient resource allocation through cell-specific configuration and load distribution.
2Ease of operation
If network reconfiguration is performed without both uplink and downlink information, then operational simplicity is maintained, but prediction accuracy deteriorates
Solution Approach 1:
The system implements feedback mechanisms where network performance measurements from both uplink and downlink are continuously collected and fed back to the optimization module. This feedback loop enables accurate prediction of reconfiguration outcomes by incorporating real-world performance data, allowing the system to learn from actual network behavior and improve prediction accuracy while maintaining automated operation.
3Measurement precision
If detailed user knowledge and specific information are fully utilized for network optimization, then network performance prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The patent introduces a predictive modeling module as an intermediary that processes detailed user knowledge and network information. This module acts as a mediator between raw data collection and optimization decision-making, transforming complex inputs into actionable predictions. The intermediary handles the complexity of data processing while presenting simplified outputs to the optimization system, thereby improving prediction accuracy without proportionally increasing overall system complexity.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
A device obtains uplink information associated with a base station and obtains downlink information associated with a mobile device in communication with the base station. The device determines observed network performance information based on the uplink information and the downlink information and determines a predictive model, based on the uplink information and the downlink information, to predict network performance information. The device also changes network configuration data, associated with the base station, to generate changed network configuration data and determines predicted network performance information for the changed network configuration data based on the predictive model. The device further selectively transmits the changed network configuration data, to the base station, based on comparing the predicted network performance information and the observed network performance information.