Trajectory Segmentation Using Curve Boundary Replacement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing trajectory segmentation methods often divide curve sections into too many segments, leading to processing delays and potential safety issues in complex traffic situations.
Innovation Solution
A method combining curve smoothing and curve detection algorithms to identify and replace segmentation points within curve segments with their start and end points, preventing unnecessary segmentation within curves and ensuring segmentation only at straight sections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional segmentation methods are used to divide trajectory data into segments of equal length or based on geometric properties, then the trajectory data can be processed and analyzed, but curve sections are divided into too many segments causing processing delays and potential safety issues
Solution Approach 1:
The patent applies different segmentation criteria to different parts of the trajectory. Straight sections are segmented into equal-length segments, while curve sections are identified and kept as single segments or fewer segments. This local differentiation prevents excessive segmentation of curves while maintaining regular segmentation of straight portions, thereby improving processing efficiency without sacrificing analysis capability.
Solution Approach 2:
The patent performs curve detection and identification before final segmentation. By first detecting which portions of the trajectory are curves (using curvature thresholds and continuous curve tracking), the system can then apply appropriate segmentation strategies to each portion. This preliminary action prevents the subsequent segmentation algorithm from creating unnecessary segments within curve sections.
2Measurement precision
If segmentation points are defined within curves based on geometric inconsistency, then trajectory points with inconsistent geometric properties are identified, but curve sections are unnecessarily divided into multiple segments
Solution Approach 1:
The patent applies different segmentation rules to different local regions of the trajectory. In curve regions (identified by curvature threshold), segmentation is suppressed or reduced. In straight regions, traditional segmentation based on geometric inconsistency is applied. This local quality approach maintains segmentation accuracy where needed while avoiding unnecessary segmentation in curve portions.
Solution Approach 2:
The patent performs preliminary curve detection by calculating curvature at each trajectory point and identifying continuous curve sections. This preliminary identification of curve regions allows the subsequent segmentation algorithm to skip or reduce segmentation operations within these regions, preventing the loss of processing time while maintaining accurate segmentation in straight sections.
3Loss of information
If more segmentation points are used to capture trajectory details, then the trajectory representation becomes more detailed, but processing delays increase and road safety may be compromised
Solution Approach 1:
The patent preserves trajectory details in straight sections through regular segmentation while reducing or eliminating segmentation in curve sections. This local differentiation maintains the necessary trajectory information for analysis without creating excessive segments that would cause processing delays and potential safety issues in real-time applications.
Solution Approach 2:
By performing preliminary curve detection and identification before segmentation, the system can anticipate which regions require detailed representation and which can be represented with fewer segments. This preliminary action ensures that trajectory details are preserved where geometrically necessary while avoiding unnecessary segmentation that would increase processing delays.
Data Source
Figure 1
Figure 2~4
Figure 5
AI summary
The invention relates to a method for segmenting trajectory data, in particular by means of a controller. Trajectory data with a plurality of points is received; the Douglas-Peucker algorithm is used on the trajectory data in order to ascertain substantially straight sections separated by segmentation points; an algorithm is used on the trajectory data in order to detect curves, in particular in the form of starting points and end points of curves; at least one segmentation point ascertained in an interval between a starting point and an end point of a curve is replaced with the starting point and the end point; and the trajectory data is segmented at the ascertained segmentation points, the starting points, and the end points. The invention additionally relates to a controller, to a computer program, and to a machine-readable storage medium.