Inferring UE Location via Sector Transition History
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional location services in wireless networks often provide inaccurate user equipment (UE) location data, especially in environments where GPS or aGPS is not available, by defaulting to the location of the radio device, which can be several kilometers away from the actual UE position, leading to poor location representation.
Innovation Solution
The solution involves inferring UE location based on historical equipment density and sector transitions, using kernel density estimation (KDE) to determine densely populated areas within sectors, and correlating this data with historical and supplemental data to improve location accuracy, even in sparse accurate location data environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional location services use shorthand techniques (radio device location, sector centroid) to indicate UE location, then location data can be provided without GPS/aGPS, but location accuracy deteriorates significantly (errors of several kilometers)
Solution Approach 1:
The patent introduces an intermediary approach by using historical accurate location data and sector transition information as a mediator between the UE and the location service. Instead of directly using imprecise shorthand techniques, the system uses historical data patterns and sector transitions to infer more accurate current locations, thereby improving measurement precision while maintaining service availability.
Solution Approach 2:
The patent applies preliminary action by collecting and storing historical accurate location data and sector transition information in advance. This historical data is then used to infer current UE locations, allowing the system to provide accurate location estimates without requiring real-time GPS or aGPS data from the UE itself.
2Measurement precision
If accurate location data (GPS, aGPS) is collected from UEs, then location accuracy improves, but data sparsity increases because relatively few UEs provide accurate location data
Solution Approach 1:
The patent applies universality by making the accurate location data and sector transition information collected from a subset of UEs serve multiple purposes. The historical data is used to infer locations for UEs that do not provide accurate location data, thereby extending the utility of the sparse accurate data to the broader UE population.
Solution Approach 2:
The patent uses copying by creating a model based on historical accurate location data and sector transition patterns. This model is then applied to infer locations for current UEs, effectively copying the spatial distribution patterns observed in historical data to predict current locations, thereby amplifying the impact of the original sparse accurate data.
3Measurement precision
If the system uses historical accurate location data and sector transitions to infer current UE locations, then location accuracy improves for UEs without GPS, but system complexity increases due to data collection and correlation requirements
Solution Approach 1:
The patent applies segmentation by dividing the location inference problem into separate components: collecting historical accurate location data, collecting sector transition information, correlating these data to determine dense bin locations, and using the correlated data to infer current UE locations. This segmentation allows each component to be optimized independently while working together to achieve accurate location estimation.
Data Source
AI summary
Determining a location of a user equipment (UE) based on historical location data and historical sector transition data is disclosed. A correlation between historic location information and a historic sector transition can be determined. The correlation can be stored in a searchable data set. A location of a current UE can be inferred based on a sector transition of the current UE. The sector transition of the current UE can be searched against eh data set to indicate a likely location of the current UE based on historical information. The searchable data set can be based on sparse location data enabling location determinations for a current UE that can otherwise lack location services. Moreover, an order of a sector transition can imbue a directionality to stored location information such that a likely location in a sector can be correlated to a transition from a prior sector of a network session of the UE.