Cellular Traffic Prediction Using Transportation and Household Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network planning for cellular sites lacks accurate methods to predict cellular traffic, particularly in mobility networks like 5G, making it difficult to select optimal locations for new sites.
Innovation Solution
A predictive model is developed using publicly available transportation data, geographic data, and census data to forecast cellular traffic by training on existing cell site data and predicting traffic at candidate sites using machine learning techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional network planning methods are used for cellular site selection, then the process is simple and quick, but the accuracy of cellular traffic prediction is insufficient
Solution Approach 1:
The system performs preliminary data collection and model training using historical cellular traffic data, transportation data, geographic data, and census data from existing cell sites. This pre-processing and training phase enables the predictive model to accurately forecast traffic at candidate sites without requiring complex real-time analysis during the site selection process itself.
Solution Approach 2:
A machine learning predictive model serves as an intermediary between raw data sources (transportation data, geographic data, census data) and decision-makers. The model processes multiple data types and transforms them into actionable traffic predictions, simplifying the complex relationship between various factors and site selection outcomes.
2Measurement precision
If comprehensive data analysis is performed for site selection, then prediction accuracy improves, but the time and resources required increase
Solution Approach 1:
The system performs comprehensive data analysis and model training in advance, storing processed results for rapid retrieval during site selection. This allows the system to provide accurate traffic predictions for multiple candidate sites quickly, without repeating the entire analysis process for each new site evaluation.
Solution Approach 2:
The predictive model creates a virtual representation of cellular traffic patterns based on historical data from existing sites. This digital twin or copy of traffic behavior allows the system to simulate and predict traffic at candidate sites without requiring physical deployment or extensive real-time monitoring, significantly reducing evaluation time.
Data Source
AI summary
Aspects of the subject disclosure may include, for example, a method in which a processing system obtains first cellular traffic data associated with an existing cell site of a communication network, first transportation data associated with a first geographic region including the existing cell site, and first household data regarding households located in the first geographic region; the first household data comprises cellular plan data, broadband plan data, demographic data and/or economic data. The processing system constructs a predictive model for cellular traffic applicable to a candidate cell site in a second geographic region, using the cellular traffic data and the first transportation data and/or the first household data. The processing system predicts second cellular traffic at the candidate cell site, using as inputs to the model second transportation data and second household data associated with the candidate cell site. Other embodiments are disclosed.