Digital Transportation Network Creation from Probe Traces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for creating digital transportation networks from uncoordinated probe traces are limited, requiring manual seeding and lacking automation for junctions, with insufficient accuracy for applications like Advanced Driver Assistance Systems (ADAS) due to positional inaccuracies from conventional Personal Navigation Devices (PNDs).
Innovation Solution
The use of statistical analysis techniques and clustering methods to automatically distinguish road intersections, determine connected road segments, and create a contiguous transportation network from probe traces collected by navigation devices, eliminating the need for manual intervention and improving accuracy by refining and processing probe data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional PND location measurements are used to generate digital transportation networks, then device complexity is reduced, but measurement precision deteriorates (positional accuracy of +/- 10 to 15 meters is insufficient for ADAS applications requiring less than 5 meters)
Solution Approach 1:
The patent combines location measurements from multiple PNDs into aggregated probe traces, merging individual low-precision measurements into a collective high-precision dataset that accurately represents transportation network geometry
Solution Approach 2:
The patent creates a digital copy (representation) of the transportation network from probe traces, generating a simplified geometric model that captures essential network features while filtering out measurement noise and inconsistencies
2Extent of automation
If manual seeding is used to initialize digital transportation network creation, then manufacturing precision is improved (accurate initial geometry), but ease of manufacture deteriorates (requires manual intervention)
Solution Approach 1:
The system performs self-service by automatically extracting transportation network geometry from probe traces without requiring manual seeding, using statistical methods to identify roads, intersections, and network topology autonomously
Solution Approach 2:
The patent performs preliminary clustering and statistical analysis on probe traces to pre-process the data into meaningful geometric patterns before network extraction, enabling automated accurate network creation without manual initialization
3Measurement precision
If uncoordinated probe traces are used directly for network creation, then ease of operation is improved (simple data collection), but measurement precision deteriorates (insufficient accuracy for junction identification)
Solution Approach 1:
The patent segments the transportation network into distinct elements (roads, intersections, junctions) by clustering probe traces and identifying geometric patterns, breaking down the complex uncoordinated data into structured network components
Solution Approach 2:
The patent transforms probe trace data by changing parameters such as clustering thresholds, statistical metrics, and geometric tolerances to convert raw uncoordinated measurements into precise network geometry suitable for ADAS applications
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In a method for creating a digital representation of a transportation network, acquired probe traces are refined based on characteristics of the transportation network. Geographic objects associated with the transportation network are identified based on the refined probe traces. A digital geographic network is built based on the refined probe traces and identified geographic objects, and the digital representation of the transportation network is created by linking the identified geographic objects in the digital geographic network.