Bi-directional Road Geometry from Sparse Probe Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for creating road geometry from probe data are ineffective when data is sparse or noisy, particularly in bi-directional road scenarios, as they require high density and uniform probe data to accurately represent road shapes and directions.
Innovation Solution
A method and apparatus that filter and process sparse probe data to determine seed points and orientations, using techniques like principal component analysis and least median squares to create bi-directional road geometry polylines, even in areas with low probe density and uncertain trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high density and high frequency probe data is used, then trajectory shape accurately approximates road geometry, but the method fails in sparse and noisy data conditions
Solution Approach 1:
The patent transforms probe data by computing bearing angles between sequential probe points and applies statistical methods (median, mean, standard deviation) to aggregate bearing information. This parameter transformation allows the system to extract road geometry signals from sparse and noisy probe data without requiring high data density, resolving the contradiction between measurement precision and data quantity requirements
Solution Approach 2:
The patent introduces bearing angles as an intermediary parameter that mediates between raw probe data and final road geometry. By computing bearing angles and using statistical aggregation (median/mean bearing, bearing standard deviation) as intermediate processing steps, the system can accurately approximate road geometry even when probe data is sparse or noisy, eliminating the need for high data density
2Adaptability or versatility
If existing road geometry techniques are used, then unidirectional geometry can be created, but bi-directional geometry cannot be accurately represented
Solution Approach 1:
The patent segments the probe data processing into distinct directional components by computing bearing angles that capture directionality information. By processing bearings separately and identifying opposing bearing patterns, the system creates separate polylines for each direction of travel, enabling bi-directional geometry representation while utilizing heading information that existing techniques lose
3Reliability
If probe data with positional and heading errors is processed, then noise is introduced, but filtering may remove valid trajectory information
Solution Approach 1:
The patent employs statistical feedback mechanisms by computing median and mean bearings along with bearing standard deviations from multiple probe points. This feedback approach allows the system to identify and correct noisy bearing measurements while preserving valid trajectory information, improving road geometry reliability without excessively filtering out valid data
Solution Approach 2:
The patent transforms noisy positional and heading error data into bearing angle parameters, then applies statistical aggregation (median, mean, standard deviation) to filter noise. This parameter transformation and statistical processing reliably distinguishes between noise and valid trajectory information, maintaining accuracy without excessive filtering
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method, apparatus and computer program product are provided to define bidirectional road geometry based upon a plurality of probe points. An orientation representative of one or more probe points is determined and a seed point representative of one or more probe points is then determined. The seed point is based on a determination of a constrained weighted center of mass of the one or more probe points. The movement of the constrained weighted center of mass is limited to a direction based upon respective headings of the one or more probe points. Polylines are created to be representative of a bi-directional road geometry. The polylines are created to be representative of one direction of travel along the bi-directional road geometry by extending a first polyline in the orientation from the seed point and by extending a second polyline in an opposite orientation from the seed point.