On-Board Vision Localization for Drift-Free UAS Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems for unmanned aerial vehicles, particularly in challenging 3D terrains like Mars, face issues with drift in position estimates due to the lack of a global reference, making precision navigation and loop closure difficult, especially in high-relief landscapes.
Innovation Solution
A vision-based localization method that uses a feature-based approach to match on-board images with pre-computed ortho-projected images and digital elevation maps, leveraging a pose prior from the on-board state estimator to guide the matching process and eliminate drift, utilizing SIFT features for accurate localization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS is used for global positioning on Earth, then precise navigation is achieved, but the system cannot be applied to Mars where no global reference exists
Solution Approach 1:
The patent creates a local map copy of the Martian terrain using orbital imagery and digital elevation maps, then matches on-board camera images against this pre-computed map to determine position. This replaces GPS by creating a portable reference system that can be used anywhere on Mars where the map covers the area.
2Area of stationary object
If feature-based matching is performed without pose prior guidance, then comprehensive terrain coverage is achieved, but computational time and processing load increase significantly
Solution Approach 1:
The patent pre-computes the map from orbital imagery and digital elevation maps before the on-board flight, organizing terrain features and elevation data in advance. During flight, the system only needs to match features against this pre-processed map rather than performing comprehensive terrain analysis, significantly reducing real-time computational requirements.
Solution Approach 2:
The patent introduces a pose prior from the on-board state estimator as an intermediary that guides the feature matching process. The pose prior provides an initial estimate of vehicle position and orientation, which constrains the search space for feature matching, reducing computational time while maintaining accurate terrain coverage.
3Ease of operation
If traditional navigation methods are used in high-relief landscapes, then simple terrain is navigated easily, but precision navigation and loop closure become difficult in challenging 3D terrains
Solution Approach 1:
The patent incorporates digital elevation maps to provide local quality information about terrain vertical structure. By matching both 2D image features and 3D elevation data, the system achieves accurate localization in challenging 3D terrains while maintaining the simplicity of automated operation. The elevation information provides additional constraints that improve precision without complicating the navigation process.
Data Source
AI summary
A method and system provide for on-board localization in a unmanned aerial system (UAS). A map image is generated (using previously acquired images) of an area that the UAS is overflying. The map image is then processed by orthorectifying, referencing the map image in a global reference frame, and generating an abstract map by detecting features and locating the features in the global reference frame. The UAS is then localized by acquiring camera images during flight, selecting a subset of the camera images as localization images, detecting on-board image features (in the localization images), mapping features from the detected on-board image features to the abstract map, deleting outliers to determine an estimated 3D pose, and refining the 3D pose. The localized UAS then used to autonomously navigate the UAS.