Mobile Network Heat Map Generation Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mobile network testing methods require multiple drive testing campaigns to generate complete and valid heat maps, necessitating significant effort and time.
Innovation Solution
A method utilizing a machine learning model, specifically a deep learning approach, to simulate participant behavior and fill missing information on existing heat maps by recognizing relationships between network data, reducing the need for extensive testing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple drive testing campaigns are conducted to generate complete heat maps, then the completeness and validity of heat map data is improved, but the time and effort required for testing increases significantly
Solution Approach 1:
The system performs preliminary actions by training the machine learning model on complete heat map data from multiple testing campaigns in advance. Once trained, the model can generate complete heat maps from single-test data, eliminating the need to repeatedly perform multiple full testing campaigns while maintaining data completeness.
Solution Approach 2:
The machine learning model learns to copy the patterns and relationships present in complete heat map data generated from multiple testing campaigns. The model replicates the completeness of multi-campaign heat maps by inferring missing information based on learned correlations, effectively creating a copy of complete heat map quality from single-test inputs.
2Loss of information
If multiple testing campaigns are run to obtain sufficient data for heat maps, then the quality and completeness of network coverage information is improved, but the complexity and cost of testing operations increases
Solution Approach 1:
The machine learning model acts as an intermediary between the test results and the heat map generation process. It processes test results and infers missing network coverage information by recognizing patterns, thereby completing the information gap without requiring multiple complex testing campaigns.
Solution Approach 2:
The system changes the parameter of testing quantity from multiple campaigns to a single test by introducing the machine learning model. The model compensates for the reduced testing quantity by using learned parameter relationships to infer missing information, maintaining information completeness while reducing operational complexity.
3Productivity
If a single test is performed to reduce testing time, then the speed of obtaining heat map results is improved, but the completeness and validity of the heat map data deteriorates
Solution Approach 1:
The system replaces the mechanical approach of performing multiple physical testing campaigns with a computational machine learning model. The model substitutes repeated field testing with algorithmic inference, maintaining data validity through learned patterns while achieving rapid single-test heat map generation.
Solution Approach 2:
The machine learning model performs preliminary learning from complete heat map data before actual heat map generation. This preliminary training enables the model to validate and complete single-test results accurately, ensuring data validity is maintained even when only a single test is performed.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
A method for mobile network testing is described, wherein at least one testing device (12) is used that is configured to be connected to the mobile network. A predefined set of test procedures is run on the at least one testing device (12) for generating at least one heat map assigned to the mobile network. The test results of the test procedures are evaluated via a machine learning model. The machine learning model is trained to complete missing information on an already present heat map based on the test results obtained. Further, a test system (10), a method for mobile network testing and a network testing system (20) are described.