Mobility Observation Using Sliding Time Intervals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobility observation techniques face challenges with sparse or high-frequency locational data, leading to inaccurate results, inefficiency, and inability to adapt to different data sources, particularly due to issues with data accuracy and distribution.
Innovation Solution
A framework that collects locational data and groups it into sliding time intervals of variable length, converting it into geometric representations such as polygons, which are stored in a multi-dimensional array to determine movement patterns, allowing for error tolerance and scalable analysis without requiring even data distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If similarity based solution is used for sparse locational data, then trajectory clustering can be performed, but accuracy deteriorates because data points for similar trajectories may be far apart
Solution Approach 1:
The patent transforms the problem from spatial proximity (2D/3D space) to temporal proximity (time dimension). Instead of measuring distance between data points in space, the system measures time intervals between consecutive locations for each trajectory. This dimensional transformation allows sparse spatial data to be effectively clustered based on temporal patterns, resolving the contradiction between handling sparse data and maintaining clustering accuracy.
2Productivity
If state based approach is used for high frequency locational data, then movement patterns can be analyzed, but the system is overwhelmed by the enormous number of states requiring compression
Solution Approach 1:
The patent extracts only the essential temporal information (time intervals between consecutive locations) from the high-frequency locational data, discarding redundant spatial details. This extraction approach reduces the data volume significantly while preserving the movement pattern information needed for analysis, thereby resolving the contradiction between maintaining analysis capability and reducing processing complexity.
3Measurement precision
If density based solution is used without normalization, then location importance can be reflected by data point density, but the density does not reflect the true distribution of moving objects due to varying time frequencies
Solution Approach 1:
The patent changes the measurement parameter from spatial density (data points per unit area) to temporal density (data points per unit time). By measuring the time intervals between consecutive locations rather than the spatial concentration of points, the system accurately reflects the true distribution and importance of locations regardless of varying sampling frequencies, resolving the contradiction between measurement precision and data adaptability.
Data Source
AI summary
Representative implementations provide devices and techniques for observing the movement of one or more objects over a duration of time. Collected locational data of the objects is grouped according to a plurality of sliding time intervals having a variable length. The data is converted to one or more geometric representations, which are representative of object movement during a time interval. Movement patterns of the objects may be determined based on the properties of the geometric representations.