POI Traffic Analysis Using Centroid Pattern Reuse
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing mobile communication device location data to infer human movement patterns are inefficient, particularly in processing short-duration travel routes like commutes, which result in high computational loads and processing times due to the need for repeated application of routing algorithms.
Innovation Solution
A method and system that utilize clustering and patterning applications to identify and store patterns in location data, allowing for the association of routes with centroids and the inference of travel modes, thereby reducing the need for repeated routing algorithm processing by recognizing and reusing previously determined patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If routing algorithms are applied repeatedly to process mobile communication device location data, then route determination accuracy is improved, but computational processing time increases
Solution Approach 1:
The system performs preliminary clustering of location data to identify centroids and pre-determines routes between these centroids using routing algorithms. These pre-computed routes are stored and reused when devices travel between the same centroids, avoiding repeated application of routing algorithms and reducing computational processing time while maintaining route determination accuracy.
Solution Approach 2:
The system performs preliminary analysis to identify repeating travel patterns in device location data. When patterns are detected, the system retrieves previously determined routes from storage instead of re-computing them, significantly reducing computational processing time for devices with predictable travel behaviors while maintaining accurate route determination.
2Loss of information
If routing algorithms are applied to every location data processing instance, then complete route information is obtained, but processing resources are consumed excessively
Solution Approach 1:
The system pre-computes and stores routes between centroids in a lookup table. When processing location data, the system first determines if the device is traveling between known centroids, and if so, retrieves the pre-computed route information instead of executing routing algorithms. This approach maintains complete route information availability while dramatically reducing processing resource consumption for repeated routing calculations.
Solution Approach 2:
The system creates and stores copies of determined routes in a database for future retrieval. Instead of re-computing routes that have been previously calculated, the system copies and reuses these stored route representations, reducing processing resource consumption while ensuring complete route information is available for analysis.
3Measurement precision
If clustering analysis is performed on all location data, then location pattern recognition is improved, but processing time increases
Solution Approach 1:
The system segments location data processing into distinct phases: clustering analysis to identify centroids, pattern recognition to detect repeating travel behaviors, and route determination. By separating these functions and storing intermediate results (centroids and patterns), the system improves location pattern recognition accuracy through comprehensive clustering while reducing overall processing time by avoiding redundant clustering operations on the same data.
Solution Approach 2:
The system performs clustering analysis preliminarily to establish centroids and travel patterns, storing these results for future use. Subsequent location data processing leverages these pre-established centroids and patterns rather than performing full clustering analysis each time, improving location pattern recognition through thorough initial clustering while reducing processing time for ongoing analysis.
Data Source
AI summary
A method of determining presence of mobile communication devices at points-of-interest. The method comprises analyzing location data of a mobile communication device by a clustering application executing on a computer system, determining clusters of the location data by the clustering application, determining cluster centroids by the clustering application, determining a plurality of routes traveled by the device by a routing application based on the centroids, analyzing the centroids and the routes by a patterning application, determining a pattern of association between the centroids and the routes by the patterning application, observing a previously determined pattern of centroids by the patterning application, and based on observing the previously determined pattern of centroids, determining a route traveled by the device by the patterning application, whereby an instance of determining the route traveled by the device by the routing application is avoided.