Spatiotemporal Causal Interaction Detection in Traffic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing location-acquisition technologies face challenges in understanding unusual spatiotemporal data from service vehicles, particularly due to data sparseness and distribution skewness, which hinders effective analysis and decision-making for traffic management.
Innovation Solution
A process that collects GPS points from service vehicles, divides geographical areas into regions, generates links based on transitions, calculates a score for minimum distortions, and identifies temporal outliers to detect abnormal traffic patterns, constructing outlier trees and determining spatiotemporal causal relationships to provide recommendations for traffic diversion and infrastructure improvements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If GPS data from service vehicles is collected and analyzed directly, then traffic pattern information is obtained, but data sparseness and distribution skewness prevent effective analysis
Solution Approach 1:
The geographical area is divided into multiple regions, and the GPS trajectory data is segmented into transitions between these regions. This segmentation transforms sparse individual vehicle paths into aggregated regional transition patterns, making the data suitable for analysis despite its sparseness.
Solution Approach 2:
Region transitions serve as an intermediary representation between raw GPS points and traffic pattern analysis. By introducing regional aggregation as an intermediate layer, the system bridges the gap between sparse individual trajectories and meaningful traffic flow patterns.
2Ease of manufacture
If the area is divided into regions and transitions are generated, then data structure is improved for analysis, but computational complexity increases
Solution Approach 1:
The continuous geographical space is segmented into discrete regions, transforming complex continuous trajectory data into simplified discrete transition counts. This segmentation reduces computational complexity by replacing continuous coordinate processing with discrete regional aggregation.
Solution Approach 2:
The system changes the parameter representation from continuous GPS coordinates to discrete regional transition counts. This parameter transformation simplifies the data structure and enables more efficient computational processing of traffic patterns.
3Measurement precision
If outlier detection is performed on raw GPS data, then abnormal patterns may be found, but false positives increase due to data sparseness
Solution Approach 1:
By segmenting individual GPS trajectories into regional transitions and aggregating these transitions, the system separates genuine traffic pattern signals from noise. The aggregation process enhances the signal-to-noise ratio, making outlier detection more accurate and reducing false positives.
Solution Approach 2:
Multiple individual vehicle trajectories are merged into aggregated regional transition data. This merging process consolidates sparse individual paths into robust collective patterns, improving the reliability of outlier detection by distinguishing true anomalies from random variations.
Data Source
AI summary
Techniques for detecting outliers in data and determining spatiotemporal causal interactions in the data are discussed. A process collects global positioning system (GPS) points in logs and identifies geographical locations to represent the area where the service vehicles travelled with a passenger. The process models traffic patterns by: partitioning the area into regions, segmenting the GPS points from the logs into time bins, and identifying the GPS points associated with transporting the passenger. The process projects the identified GPS points onto the regions to construct links connecting GPS points located in two or more regions. Furthermore, the process builds a three-dimensional unit cube to represent features of each link. The points farthest away from a center of data cluster are detected as outliers, which represent abnormal traffic patterns. The process constructs outlier trees to evaluate relationships of the outliers and determines the spatiotemporal causal interactions in the data.