Wireless User Geolocation Using Cell-Specific ML Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wireless service providers lack accurate methods to geolocate a significant portion of their users due to the absence of GPS data, necessitating alternative techniques for real-time location prediction.
Innovation Solution
Utilizing machine learning models trained on key performance indicators such as signal strength and timing events to predict the locations of wireless network users, with adjustable accuracy and latency based on computational constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS data is used for geolocation, then measurement precision is improved, but adaptability deteriorates because a significant portion of users lack GPS data
Solution Approach 1:
The patent introduces machine learning models as intermediaries that translate wireless network measurements (signal strength, timing events) into location predictions. These models serve as a mediator between the available network data and the desired location information, enabling geolocation for users without GPS while maintaining reasonable accuracy through learned patterns from training data
Solution Approach 2:
The patent changes the input parameters from GPS coordinates to wireless network measurements such as signal strength and timing events. By transforming the geolocation problem to use different measurable parameters that are universally available across all wireless devices, the system achieves both broad user coverage and functional geolocation capability
2Adaptability or versatility
If machine learning models are used to predict locations, then adaptability is improved to cover all users, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by training machine learning models offline using historical data before deployment. The complex computational work of learning patterns from大量 data is performed in advance, allowing the deployed models to make fast predictions with minimal real-time computational overhead, thus reducing the burden on live network operations
Solution Approach 2:
The patent segments the geolocation task into distinct components: data collection from the network, offline model training, and real-time prediction execution. This segmentation allows each component to be optimized independently, with the heavy computational lifting done during the training phase rather than during live operations
3Productivity
If real-time location prediction is performed for millions of users, then productivity is improved, but use of energy increases
Solution Approach 1:
The patent employs periodic action by processing user location predictions in batches or at scheduled intervals rather than continuously for every event. This allows the system to handle millions of users efficiently by grouping computations, reducing peak energy consumption while maintaining high overall productivity through optimized batch processing
Data Source
AI summary
A method includes selecting a first machine learning model from a plurality of machine learning models that are trained for use in performing geolocation, wherein the first machine learning model is selected to perform geolocation within a first cell of a plurality of cells of a wireless network, acquiring event data from a plurality of wireless devices within the first cell, grouping the event data into a plurality of records, wherein each record of the plurality of records contains event data that indicates a common wireless device of the plurality of wireless devices, a common cell of the plurality of cells, and a common timestamp, and generating a predicted location of a first wireless device of the plurality of wireless devices, using the first machine learning model, wherein the first machine learning model outputs the predicted location in response to an input of a record of the plurality of records.