Movement Line Information Generation for Dense Tracking Ambiguity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for confirming the association between a trajectory and identification information of a moving body are prone to errors due to increased calculation costs and ambiguity, especially in dense tracking regions, leading to the discarding of true hypotheses and unreliable associations.
Innovation Solution
A movement line information generation system that includes a trajectory link candidate generation mechanism, likelihood calculation, identification information association, association trend quantification, and association score calculation to estimate and quantify the ambiguity of associations between trajectories and identification information, even in situations where accurate associations cannot be specified.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the number of moving bodies increases, then the tracking coverage is improved, but the calculation cost increases
Solution Approach 1:
The patent segments the trajectory into multiple trajectory candidates and divides the hypothesis generation process into stages. Instead of generating all hypotheses at once, the system processes trajectory segments sequentially, reducing the computational burden of handling all moving bodies simultaneously.
Solution Approach 2:
The patent performs preliminary processing by generating trajectory candidates and segments before hypothesis generation. This preliminary action organizes the data structure in advance, allowing for more efficient hypothesis generation and reducing the overall calculation cost when tracking multiple moving bodies.
2Loss of energy
If pruning processing is performed to restrict calculation cost increase, then the calculation cost is reduced, but true hypotheses may be discarded
Solution Approach 1:
The patent performs preliminary generation of comprehensive trajectory candidates and segments before hypothesis generation. This ensures that all possible true hypotheses are captured in the candidate set before any filtering occurs, preventing true hypotheses from being discarded during pruning.
Solution Approach 2:
The patent dynamically adjusts the hypothesis generation process by evaluating likelihoods for each trajectory candidate and identification information pair. This dynamic approach allows the system to retain hypotheses that meet reliability thresholds while discarding only those with sufficiently low likelihoods, balancing calculation cost and hypothesis accuracy.
3Ease of operation
If maximum-likelihood hypothesis is selected, then the association is simplified, but the true association may not be identified
Solution Approach 1:
The patent calculates likelihoods for each trajectory link/identification information pair and uses this feedback to iteratively improve the association. Rather than a single maximum-likelihood selection, the system uses likelihood calculations to guide the association process, allowing for correction and refinement of associations based on accumulated evidence.
Solution Approach 2:
The patent implements a dynamic association process where the system evaluates multiple trajectory candidates and adjusts associations based on likelihood calculations. This dynamic approach allows the system to adapt to changing conditions and identify true associations even when the maximum-likelihood hypothesis alone would be insufficient.
Data Source
AI summary
At least one processor inputs trajectories of moving bodies in a tracking region. The at least one processor generates a score corresponding to each of the identification information, the score indicating a likelihood that the moving bodies corresponding to the identification information present in the area. The at least one processor determines an area with a high density of the moving bodies that a density of the moving bodies is a predetermined value or higher. The at least one processor associate, based on the score, a set of moving bodies and a set of identification information of the moving bodies. The at least one processor controls a display to display, for the area with the high density of the moving bodies, the set of moving bodies and the set of identification information in association with the score.