Map-Based Vision Navigation Fault Isolation for Vehicle State Integrity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation systems in GNSS denied environments lack effective fault detection and integrity monitoring, particularly for vision-based sensors, which are critical for ensuring the accuracy and reliability of vehicle kinematic state estimation.
Innovation Solution
A multi-level filter system with overlapping vision sensors and inertial sensors, utilizing position probability density function filters and consistency monitors to detect and isolate faults, ensuring statistical consistency and integrity of navigation measurements by correlating image data with terrain maps and inertial data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a hybrid solution separation technique is used for integrity monitoring, then fault detection capability is improved, but the system cannot adequately handle vision-based sensors in safety critical navigation
Solution Approach 1:
The monitoring system is divided into multiple independent monitoring channels, each processing measurements from specific sensor subsets. This segmentation allows the system to adapt to different sensor types including vision-based sensors while maintaining robust fault detection capabilities through parallel independent analysis paths.
Solution Approach 2:
The monitoring architecture is designed to be sensor-type agnostic, capable of processing measurements from various sensor modalities including vision-based sensors, inertial sensors, and navigation sensors. The universal framework applies consistent integrity monitoring principles across different sensor types, enabling adaptability while maintaining reliability.
2Reliability
If multiple vision sensors with overlapping FOV are used, then measurement redundancy is improved, but device complexity increases
Solution Approach 1:
The sensor system is organized into distinct monitoring channels, with each channel processing a specific subset of sensors and their measurements independently. This segmentation manages complexity by creating modular processing paths while the combined output from multiple channels provides enhanced measurement redundancy and cross-validation.
Solution Approach 2:
A centralized monitoring processor acts as an intermediary that receives and coordinates data from multiple vision sensors with overlapping fields of view. This intermediary manages the complexity of fusing redundant measurements while extracting consistent navigation information, transforming complex multi-sensor inputs into reliable estimated states.
Data Source
AI summary
A system for integrity monitoring comprises a processor onboard a vehicle, a terrain map database, vision sensors, and inertial sensors. Each vision sensor has an overlapping FOV with at least one other vision sensor. The processor performs integrity monitoring of kinematic states of the vehicle, and comprises a plurality of first level filter sets, each including multiple position PDF filters. A consistency monitor is coupled to each first level filter set. Second level filter sets, each including multiple position PDF filters and a navigation filter, are coupled to the consistency monitor, the map database, the vision sensors, and the inertial sensors. Consistency check modules are each coupled to a respective second level filter set. A third level fusion filter is coupled to each of the consistency check modules. An integrity monitor is coupled to the fusion filter and assures integrity of a final estimate of the vehicle kinematic state statistics.