Ground-Facing Camera Localization for GPS-Denied Aerial Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-flying aerial systems face challenges in GPS-denied environments due to reliance on external sources, making them vulnerable to interference and failure, especially in contested or extreme conditions where visual cues are unreliable.
Innovation Solution
A neural network-based system using a ground-facing camera and a twin network architecture is trained with extensive pre-flight imagery data to match in-flight images with a 3D surface model, enabling robust navigation without frequent updates or cloud cover, utilizing multiple spectral regions for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS is used for navigation, then navigation accuracy is improved, but reliability deteriorates in GPS-denied environments due to external interference and signal jamming
Solution Approach 1:
The patent introduces an intermediary system consisting of a ground-facing camera and a trained neural network model that acts as a mediator between the aerial system and the terrain below. This intermediary enables the system to determine position by capturing images of ground features and matching them against a pre-trained model, thereby providing navigation capability independent of GPS satellites and resistant to signal jamming.
2Reliability
If downward facing camera with 3D surface model matching is used, then GPS-denied navigation is enabled, but the system requires frequent database updates and clear weather conditions
Solution Approach 1:
The patent applies preliminary action by pre-training the neural network model with extensive terrain imagery data covering diverse weather conditions, seasonal variations, and different terrain types before the aerial mission. This pre-training enables the system to adapt to varying weather conditions and terrain types during operation without requiring real-time database updates or clear weather conditions, as the model has already learned to recognize features across these variations.
3Reliability
If high-altitude flight is used, then line of sight to satellites is improved, but visual cues for navigation become too distant and unreliable
Solution Approach 1:
The patent applies dimensionality change by shifting the navigation approach from relying on distant visual cues in the horizontal plane to using a ground-facing camera that captures vertical imagery of the terrain directly below. This dimensional shift allows the system to use the trained neural network model to recognize and match terrain features even at high altitudes, effectively creating a new dimension for navigation that is not constrained by the distance to visual cues.
Data Source
AI summary
A high-altitude platform and method for in-flight navigation, the platform comprising a ground-facing camera and a computing device including a memory and a processor, the memory storing instructions which when processed cause the computing device to perform a position determination method, comprising: acquiring one or more images from the ground-facing imaging device; retrieving, from an image dataset associated with a trained model, a plurality images associated with an area of interest; processing, using an image comparator, the acquired one or more images and the retrieved images associated with terrain of interest to derive therefrom respective acquired image feature sets; using a loss function, comparing each acquired image feature set to each retrieved image feature set to identify for each acquired image a corresponding matching retrieved image; using the matched image and the optical characteristics of the ground-facing camera, determining the position of the platform.