Traffic Information Device Feature Space Link Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic information systems face challenges in accurately estimating and interpolating missing data from probe vehicles, especially when the rate of missing data is high, leading to unreliable predictions and reduced coverage area due to the inability to differentiate between links with low and high correlation.
Innovation Solution
A traffic information providing device that generates a feature space based on historical data to analyze correlation among links, performs inverse projection to estimate current traffic information, and filters links with low correlation to prioritize data collection and interpolation, ensuring accurate interpolation and expanding coverage area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If data interpolating technique is used to estimate missing traffic information, then coverage area is expanded, but measurement precision deteriorates when missing data rate is high
Solution Approach 1:
The patent segments the feature space into multiple subspaces based on the degree of correlation among links. By performing interpolation in each subspace separately rather than in the entire feature space, the system maintains higher precision even when missing data rate is high, while still expanding coverage area.
Solution Approach 2:
The patent applies different interpolation strategies to different regions of the feature space based on local correlation characteristics. Links with high correlation are interpolated using standard methods, while links with low correlation are handled differently, ensuring local precision is maintained throughout the coverage area.
2Measurement precision
If projection method is used to interpolate missing data by decomposing traffic information, then interpolation accuracy is improved, but device complexity increases
Solution Approach 1:
The patent reduces computational complexity by segmenting the feature space into subspaces and performing interpolation only in necessary subspaces. This avoids the need to process the entire feature space, thereby reducing device complexity while maintaining interpolation accuracy through targeted processing.
Solution Approach 2:
The patent applies partial action by performing interpolation operations only on the necessary subspaces rather than the complete feature space. This selective approach reduces computational burden and device complexity while achieving sufficient interpolation accuracy for the required coverage area.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A probe center server (10) previously performs filtering to determine whether or not a projection norm of a link in a feature space can be interpolated, and notifies a probe terminal (20) to preferentially collect and upload detected probe data for the link whose missing data can not be interpolated.