Hybridization Device Kalman Filter Voting for INS/GNSS Error Minimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation systems using hybrid INS/GNSS face issues with long-term inertial unit degradation and noisy satellite navigation data, where existing architectures fail to ensure that corrections applied to the virtual platform are not polluted by satellite failures, leading to potential errors in position calculation.
Innovation Solution
A hybridization device with a bank of Kalman filters and a correction generation module that analyzes the signs of correction vectors from each filter to generate a stabilization vector component-wise, ensuring that only consistent corrections are applied to the virtual platform, thereby minimizing errors and avoiding pollution from satellite failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If corrections from a single selected Kalman filter are applied to the virtual platform, then the system complexity is reduced, but the reliability deteriorates because the correction may be polluted by undetected satellite failures
Solution Approach 1:
The patent segments the correction vector generation into multiple independent Kalman filters, each producing a correction vector. Instead of relying on a single filter, the system divides the correction task across multiple filters and combines their outputs through a voting mechanism, thereby reducing the impact of any single failed filter while maintaining system functionality.
Solution Approach 2:
The patent implements a feedback mechanism where the correction vectors from multiple Kalman filters are continuously monitored and compared. The voting mechanism provides feedback on the consistency of corrections, and when inconsistencies are detected (indicating potential satellite failures), the system adjusts which filters contribute to the final correction, ensuring reliable operation.
2Reliability
If a bank of Kalman filters is used to detect satellite failures, then the reliability of correction selection improves, but the device complexity increases due to multiple filters and selection logic
Solution Approach 1:
The patent divides the correction computation into multiple independent Kalman filters, each processing satellite data differently (e.g., excluding different satellites). This segmentation enables failure detection through comparison while keeping each individual filter relatively simple and modular.
Solution Approach 2:
The patent merges the outputs of multiple Kalman filters through a voting mechanism to produce the final correction vector. This combining approach distributes the reliability function across multiple filters while using a simple majority-vote logic to integrate their results, avoiding the need for complex selection algorithms.
3Ease of operation
If the virtual platform is stabilized using corrections from a single filter, then the ease of operation improves, but the measurement precision deteriorates because polluted corrections cannot be distinguished from valid ones
Solution Approach 1:
The patent segments the correction vector into multiple candidate corrections from different Kalman filters. By maintaining these segmented corrections separately and applying a voting mechanism, the system preserves precision through comparison while keeping the final application simple through automated majority voting.
Solution Approach 2:
The voting mechanism provides continuous feedback on the consistency of correction vectors from different filters. This feedback automatically identifies and excludes polluted corrections without requiring complex manual intervention, thereby maintaining both precision and ease of operation.
4Loss of information
If reconfiguration by recopying data from an unpolluted filter is performed, then the loss of information is reduced, but the reliability deteriorates because the excluded satellite may not be the one containing the fault
Solution Approach 1:
The patent maintains segmented correction vectors from multiple Kalman filters simultaneously, each excluding different satellites. This segmentation ensures that if one filter is polluted, others remain valid, preventing information loss while avoiding the need to identify which specific satellite is faulty.
Solution Approach 2:
The patent merges corrections from multiple filters through voting, which automatically combines information from all filters while excluding polluted ones. This approach recovers navigation accuracy without requiring precise fault isolation, as the voting mechanism naturally discounts inconsistent (polluted) corrections.
Data Source
Figure 1
AI summary
The invention relates according to a first aspect to a hybridization device (1) comprising a virtual platform (2), a bank (3) of Kalman filters each estimating a correction vector (dXO-dXn) comprising a plurality of components, said device formulating a hybrid output (SH) corresponding to inertial measurements (PPVI) calculated by the virtual platform (2) and corrected by a stabilization vector (dC) exhibiting one and the same plurality of components, characterized in that it comprises a correction formulation module (4) configured so as to formulate each of the components (dC[state]) of the stabilization vector (dC) as a function of all the corresponding components (dXO[state]-dXn[state]) of the correction vectors (dXO-dXn).