Behavior Pattern Extraction Using Multi-Source Location Data Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing behavior pattern extraction systems face challenges in accurately extracting user behavior patterns due to uncertainties in GPS location information, especially when terminals are still or located underground, and variations in signal reception levels, leading to low accuracy in determining user movement and stay points.
Innovation Solution
A behavior pattern extraction system that measures location points, sets staying points and records based on concentrated location points, extracts representative staying points with high accuracy by considering error ranges, and records these as behavior pattern information to improve accuracy in user behavior analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS location information is used to extract behavior patterns, then location-based behavior analysis can be provided, but the accuracy of determining user movement and stay points deteriorates due to GPS uncertainty and signal variations
Solution Approach 1:
The patent combines multiple location information acquisition methods (GPS, base station triangulation, Wi-Fi positioning) to obtain location data from multiple sources. By merging these different positioning methods, the system compensates for the weaknesses of individual methods and achieves more reliable behavior pattern extraction even when GPS accuracy is poor.
Solution Approach 2:
The system performs preliminary processing of location data by calculating not only location coordinates but also movement speed, acceleration, and time stamps before behavior pattern extraction. This preliminary action prepares refined data that accounts for GPS uncertainty, enabling more accurate determination of stay points and movement patterns in subsequent processing.
2Productivity
If location information is acquired periodically from GPS, then continuous location tracking is achieved, but the ability to accurately judge whether the terminal stands still deteriorates due to GPS calculation errors
Solution Approach 1:
The patent dynamically adjusts the determination criteria for stay points based on accumulated location data over time. Instead of using fixed thresholds, the system calculates movement speed and acceleration from sequential location points and adapts the stationary judgment criteria according to the observed dynamics, enabling accurate distinction between genuine stationary states and GPS error-induced position variations.
Solution Approach 2:
The system uses feedback from multiple consecutive location measurements to refine the judgment of whether the terminal is stationary. By continuously monitoring location changes and comparing them against calculated movement thresholds, the system adjusts its determination of stationary states based on accumulated evidence, improving accuracy over time despite GPS uncertainties.
3Reliability
If location points are processed to identify stay points, then behavior pattern extraction is enabled, but the accuracy of stay point identification deteriorates when location points are scattered due to environmental factors
Solution Approach 1:
The patent creates a virtual representation of physical stay points by calculating representative coordinates from scattered location measurements. Instead of directly using raw GPS coordinates, the system generates corrected location data that copies the essential information while eliminating error-induced scatter, producing stable and accurate stay point identifiers for behavior pattern extraction.
Solution Approach 2:
The system transforms the raw location data by changing parameters such as calculating movement speed, acceleration, and time intervals between location points. These parameter transformations convert scattered position data into structured movement patterns, enabling accurate identification of stay points even when original location coordinates vary due to environmental factors.
Data Source
AI summary
There is provided a behavior pattern extraction system which can extract user's behavior pattern with high accuracy. The behavior pattern extraction system includes: a location information acquiring section measuring location points which each indicate a location of a user; a staying point extraction section setting staying points and staying records based on an range in which location points which are each measured in a first period are concentrated, wherein the staying points each indicate a location where the user has stayed temporarily; a representative staying point extraction section setting a representative staying point and a representative staying record based on an range in which the staying points which are set in a second period which is longer than the first period are concentrated, wherein the representative staying point indicates a location where the user has repeatedly visited, and the representative staying record indicates an error range of the location of the representative staying point; and a behavior pattern record section recording the representative staying point and the representative staying record in a storage area as behavior pattern information of the user.