Adaptive Gate Value for Multi-Sensor Observation Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-sensor tracking systems, accurately associating observations from different sensor systems is challenging due to biases, random errors, and false or missed observations, making it difficult to overlay data from one system with another.
Innovation Solution
A method using a likelihood function, specifically the Global Nearest Pattern (GNP) technique, to assign pairs of observations from different sensor systems, accounting for random and correlated bias components, and determining a practical gate value using nearest neighbor distances to estimate the true volume.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a fixed gate value is used for associating observations from multiple sensor systems, then the association process is simplified, but the accuracy deteriorates due to biases and random errors in different sensor systems
Solution Approach 1:
The gate value is transformed from a fixed parameter to a dynamic one that adapts to each observation pair. The invention calculates individual gate values based on the specific characteristics of each sensor pair and observation combination, allowing the system to optimize association accuracy for each case rather than using a uniform threshold throughout.
Solution Approach 2:
The invention changes the gate value parameter dynamically by calculating it from the Mahalanobis distance of the nearest neighbor observation. This parameter transformation allows the gate value to reflect the actual statistical properties of each sensor system and observation pair, thereby improving measurement precision while maintaining operational feasibility through automated calculation.
2Productivity
If data from different sensor systems are directly overlaid without correction, then the processing speed is maintained, but the reliability deteriorates due to misalignment and systematic errors
Solution Approach 1:
The invention performs preliminary calculations of the gate value using the Mahalanobis distance before the actual association decision is made. This preliminary action prepares the adaptive threshold in advance, allowing the system to quickly compare observations against the pre-calculated gate value and maintain processing speed while improving reliability through accurate, error-corrected associations.
3Device complexity
If a simple association threshold is used, then the computational complexity is reduced, but the measurement precision deteriorates in the presence of false and missed observations
Solution Approach 1:
The gate value calculation method is self-adapting to the specific characteristics of each sensor system and observation set. By automatically calculating the Mahalanobis distance and deriving the gate value from the nearest neighbor, the system serves itself to determine the optimal threshold without requiring external calibration or complex manual configuration, thereby maintaining low operational complexity while achieving high precision.
Data Source
AI summary
In one aspect, a method to assign observations includes receiving first observations of a first sensor system, receiving second observations of a second sensor system and assigning a set of pairs of the first and second observations predicted to correspond to the same physical position. The assigning includes using a likelihood function that specifies a likelihood for each assigned pair. The likelihood is dependent on the assignment of any other assigned pairs in the set of assigned pairs. The assigning also includes determining the set of assigned pairs for the first and second observations based on the likelihood function. The likelihood function uses a gate value determined from estimating a true volume using nearest neighbor distances determined from the first and second observations.