Trajectory Outlier Detection Using Spatiotemporal Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing TRAOD algorithm for hurricane trajectory outlier detection does not accurately match actual situations due to its focus on spatial changes while ignoring temporal changes, leading to ineffective outlier detection.
Innovation Solution
A method that extracts characteristic points based on both spatial and temporal states of trajectory points, segments the trajectory into segments using these points, and compares them to normal and abnormal segments using spatial-temporal state distances to identify outliers, incorporating a two-stage detection process for enhanced accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the classic TRAOD algorithm uses distance-based outlier detection method, then the algorithm can detect trajectory outliers, but the detection accuracy does not match actual situations because temporal changes are ignored
Solution Approach 1:
The patent transitions from purely spatial distance measurement to spatiotemporal distance measurement by adding the time dimension. The distance function is extended to include temporal distance components (dt1, dt2, dt3) in addition to spatial components (d⊥, d∥, dθ), enabling the algorithm to consider both spatial and temporal changes when detecting trajectory outliers.
Solution Approach 2:
The patent modifies the distance calculation parameters by introducing temporal parameters (time differences dt1, dt2, dt3) alongside spatial parameters. The weight coefficients (w1, w2, w3) are applied to different temporal distance components, allowing dynamic adjustment of temporal factor importance in outlier detection.
2Productivity
If trajectory data is processed using the TRAOD algorithm, then outlier detection can be performed, but operation data volume is large reducing efficiency
Solution Approach 1:
The patent divides the trajectory into multiple segments based on characteristic points (turning points, extreme points, intersection points). By segmenting the trajectory, the algorithm processes smaller sub-trajectories independently, reducing the overall data volume that needs to be processed while maintaining detection accuracy.
Solution Approach 2:
The patent extracts only the essential characteristic points (turning points, extreme points, intersection points) from the complete trajectory data. This extraction process filters out redundant information and retains only the critical features needed for outlier detection, significantly reducing operation data volume.
3Reliability
If trajectory segmentation is performed without considering temporal states, then segmentation can be completed quickly, but local and global features are not effectively maintained
Solution Approach 1:
The patent performs preliminary identification of characteristic points (turning points, extreme points, intersection points) before conducting the main outlier detection process. This preliminary action prepares the data by marking important segments in advance, ensuring that local and global features are preserved while simplifying subsequent processing steps.
Data Source
AI summary
Disclosed is a method and a device of trajectory outlier detection. The method may include: points on a trajectory to be detected are obtained by sampling the trajectory; characteristic points are extracted from the points according to spatial state and temporal state of each of the points; trajectory segments are obtained by segmenting the trajectory according to the characteristic points; each of the trajectory segments is compared to normal trajectory segments and abnormal trajectory segments; and one or more trajectory outliers are identified from the trajectory segments based on comparison results. Wherein, the normal trajectory segments and the abnormal trajectory segments are obtained by clustering trajectory segments in a training set; and the trajectory segments in the training set are obtained by segmenting historical trajectories based on characteristic points extracted from points on the historical trajectories according to spatial state and temporal state of the points.