Cellular Edge Device Localization Without GPS Using ML Confidence Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for determining the location of mobile edge devices are resource-intensive, power-consuming, and unreliable due to the need for GPS, which can fail in adverse conditions or indoor environments, leading to high costs and inefficiencies for telecommunications network providers.
Innovation Solution
A device edge controller uses a machine learning model to determine location by combining edge parameters, geographic data, and real-time metadata, discarding unfit locations, and ranking based on confidence to select the actual location, eliminating the need for GPS and reducing power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS is used to determine device location, then location accuracy is improved, but power consumption and device complexity increase
Solution Approach 1:
The patent replaces the mechanical/GPS-based location determination system with a network-based system that uses cellular network parameters (signal strength, cell ID, handover information) combined with machine learning algorithms to infer device location, thereby eliminating the need for GPS hardware and reducing power consumption
Solution Approach 2:
The patent introduces network parameters and machine learning models as intermediaries between the device and location determination. Instead of directly using GPS, the system uses cellular network data (signal strength, cell identifiers, handover records) as intermediate measurements to infer location through trained models
2Measurement precision
If GPS is used to determine device location, then location accuracy is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces the mechanical/GPS-based location determination system with a network-based system that uses cellular network parameters (signal strength, cell ID, handover information) combined with machine learning algorithms to infer device location, thereby eliminating the need for GPS hardware and reducing power consumption
Solution Approach 2:
The patent extracts the location determination function from the device itself and relocates it to the network side. The device simply provides network parameters, while the network uses machine learning models to perform location inference, thereby removing GPS hardware requirements from the device
3Measurement precision
If GPS is used to determine device location, then location accuracy is improved, but reliability decreases in adverse conditions
Solution Approach 1:
The patent introduces network parameters and machine learning models as intermediaries between the device and location determination. Instead of directly using GPS, the system uses cellular network data (signal strength, cell identifiers, handover records) as intermediate measurements to infer location through trained models
Solution Approach 2:
The patent changes the parameters used for location determination from GPS-based (satellite signals requiring line-of-sight) to network-based parameters (signal strength, cell ID, handover information) that can be obtained in indoor and adverse conditions where GPS fails
4Use of energy by moving object
If network-based location determination is used, then power consumption is reduced, but location accuracy decreases
Solution Approach 1:
The patent changes the parameters used for location determination from GPS-based (satellite signals requiring line-of-sight) to network-based parameters (signal strength, cell ID, handover information) that can be obtained in indoor and adverse conditions where GPS fails
Solution Approach 2:
The patent applies different location determination strategies based on the local environment and available data. The machine learning model selectively uses different network parameters (signal strength, cell ID, handover information) depending on the local conditions, optimizing accuracy for each specific scenario
5Measurement precision
If multiple data sources are combined for location determination, then location accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-training machine learning models with extensive training data that captures various scenarios and conditions. This pre-training handles the complexity of processing multiple data sources, allowing the models to automatically learn which features are most important for accurate location determination in different contexts
Data Source
AI summary
A device may receive edge parameters, geographic data, traffic data, and real-time metadata associated with an approximate location of a mobile edge device and a device edge. The device may receive a request for an actual location of the mobile edge device and the device edge, and may process the edge parameters, the geographic data, the traffic data, and the real-time metadata, with a machine learning model, to calculate multiple locations. The device may discard locations that fail to fit the edge parameters, the geographic data, the traffic data, and the real-time metadata, to generate a set of locations, and may select, from the set of locations, a location with a greatest location confidence determination as the actual location of the mobile edge device and the device edge.