Vision Navigation Error Overbounding via Candidate Pose Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vision-based navigation systems face challenges in providing high-confidence error bounding for optical pose estimation in safety-critical applications, particularly due to ambiguity in correspondence between 2D image features and 3D constellation features, leading to potential system integrity issues and inaccurate pose estimates.
Innovation Solution
A vision-based navigation system that includes a camera and processors to capture 2D images of a target environment, determines candidate correspondences, evaluates viability thresholds, and calculates conditional and containment pose error bounds, using auxiliary measurement data to eliminate non-viable solutions and ensure high-confidence error overbounding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large set of candidate CMAPs is considered to ensure coverage of all possible correspondences, then the scope of candidate pose solutions is broadened, but precise high confidence error bounding becomes precluded and system integrity deteriorates
Solution Approach 1:
The patent segments the large set of candidate CMAPs into multiple groups based on residual monitoring results. By dividing the candidate set and applying different evaluation criteria to different segments, the system can maintain comprehensive coverage while enabling precise error bounding for each segment, thus resolving the contradiction between broad coverage and precision.
Solution Approach 2:
The patent applies preliminary residual monitoring to filter and rank candidate CMAPs before final pose estimation. This preliminary action identifies high-confidence candidates early in the process, allowing the system to focus computational resources on a reduced set of promising candidates while maintaining confidence in the error bounds.
2Adaptability or versatility
If residual monitoring thresholds are loosened to accept more candidate CMAPs, then more candidate pose solutions are available, but erroneous or infeasible CMAPs may be found valid leading to system integrity issues
Solution Approach 1:
The patent implements a feedback mechanism where residual monitoring results are used to iteratively refine the selection of candidate CMAPs. The system continuously evaluates candidates against residual thresholds and uses this feedback to eliminate erroneous CMAPs while retaining valid ones, thus maintaining system integrity while preserving a sufficient number of candidate solutions.
Solution Approach 2:
The patent replaces simple threshold-based filtering with a more sophisticated evaluation mechanism that combines residual monitoring with geometric consistency checks and physical plausibility criteria. This substitution allows the system to maintain reliability while accepting a broader range of candidates that would otherwise be rejected by rigid thresholding.
3Reliability
If residual monitoring thresholds are tightened to improve system integrity, then fewer erroneous CMAPs are accepted, but valid CMAPs may be discarded leading to loss of system integrity
Solution Approach 1:
The patent implements dynamic threshold adjustment based on the distribution of residual values across candidate CMAPs. Rather than using fixed thresholds, the system adapts the criteria for accepting candidates based on the specific characteristics of each evaluation context, allowing valid CMAPs to be retained while still maintaining high integrity standards. This dynamic approach prevents premature discarding of valid solutions.
Solution Approach 2:
The patent changes the parameters used for evaluating candidate CMAPs by incorporating multiple criteria beyond simple residual thresholds, including geometric consistency measures and physical plausibility checks. This multi-parameter evaluation approach allows the system to maintain strict integrity standards while preserving availability by accepting candidates that meet multiple criteria rather than a single rigid threshold.
4Measurement precision
If model-based constraints are applied to determine pose from correspondences, then pose estimation accuracy is improved, but the system assumes the correspondence set is correct rather than identifying correspondence ambiguities
Solution Approach 1:
The patent inverts the traditional approach by first identifying correspondence ambiguities and evaluating multiple candidate CMAPs before applying model-based pose estimation. Rather than assuming correspondences are correct and estimating pose directly, the system first works backwards to identify which correspondences are ambiguous and generates multiple candidate solutions, then selects the most reliable ones. This inversion allows the system to maintain pose estimation accuracy while preserving awareness of correspondence uncertainties.
Data Source
AI summary
A system and method for high-confidence error overbounding of multiple optical pose solutions receives a set of candidate correspondences between 2D image features captured by an aircraft camera and 3D constellation features including at least one ambiguous correspondence. A candidate estimate of the optical pose of the camera is determined for each of a set of candidate correspondence maps (CMAP), each CMAP resolving the ambiguities differently. Each candidate pose estimate is evaluated for viability and any non-viable estimates eliminated. An individual error bound is determined for each viable candidate pose estimate and CMAP, and based on the set of individual error bounds a multiple-pose containment error bound is determined, bounding with high confidence the set of candidate CMAPs and multiple pose estimates where at least one is correct. The containment error bound may be evaluated for accuracy as required for flight operations performed by aircraft-based instruments and systems.