Runway Vision Pose Estimation for Autonomous Aircraft Landing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current aircraft landing systems lack sufficient accuracy and precision for attitude determination, and are often unreliable due to limitations in global navigation satellite systems, inertial navigation, and ground-based augmentation systems, especially in adverse weather conditions or without infrastructure support.
Innovation Solution
A computer-vision-based system that uses onboard cameras to capture images of the runway, compares them with generated images from airfield model data to determine the aircraft's pose, adjusting altitude, bearing, and distance, and outputs refined pose estimates for flight control, independent of other infrastructure and insensitive to RF system degradations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing navigation systems (GNSS, INS) are used for autonomous landing, then basic position information can be obtained, but the accuracy and precision of final position and attitude determination is insufficient
Solution Approach 1:
The patent replaces radio frequency-based navigation systems (GNSS, INS, ILS) with an optical-based computer vision system. The aircraft uses onboard cameras to capture images of the runway and surrounding environment, then uses image processing and feature matching algorithms to determine position and attitude. This substitution eliminates susceptibility to RF signal multipath reflections and availability issues while providing sufficient measurement precision for autonomous landing.
2Extent of automation
If automatic landing systems (ALS) are installed on aircraft, then supervised automated landing can be achieved, but the cost and weight increase significantly
Solution Approach 1:
The patent replaces expensive, heavy dedicated automatic landing system hardware with commercially available onboard cameras and standard computer vision processing. Instead of using specialized ALS equipment designed for decades of service, the system uses modern imaging sensors and algorithms that can be implemented on existing aircraft without significant weight penalty, achieving automated landing capability at much lower cost and weight.
3Measurement precision
If instrument landing systems (ILS) are used to assist landing, then bearing and glideslope information can be obtained, but visual check by pilot is still required and system availability is limited
Solution Approach 1:
The patent enables the aircraft to determine its own position and attitude using its own onboard cameras and image processing capabilities, without requiring external ground-based ILS infrastructure. The system processes images of the runway and environment captured by the aircraft itself, performing self-service navigation and landing guidance that is independent of ground-based system availability.
4Reliability
If ground-based augmentation system (GBAS) is deployed, then differential corrections and integrity monitoring can be provided, but the system is only available at facilities that have purchased and installed it
Solution Approach 1:
The patent replaces infrastructure-dependent GBAS with a self-contained computer vision system that determines position and attitude through image processing of runway features. This substitution eliminates the need for ground-based augmentation infrastructure entirely, providing reliable and accurate navigation guidance at any airport with visual features that can be captured by the aircraft's cameras, thereby achieving universal geographic availability.
Data Source
Figure 1
Figure 2A~2B
Figure 2C
AI summary
An apparatus for supporting an aircraft approaching an airfield runway of an airfield is provided. The apparatus obtains a current pose estimate of the aircraft relative to the airfield runway and determines, based on the current pose estimate of the aircraft, a plurality of proximate poses of the aircraft. The apparatus generates a plurality of images of the airfield runway from a plurality of expected points-of-view of the aircraft and acquires a real-time image of the airfield runway. The apparatus performs a comparison of the real-time image and the plurality of images to identify a best-match image from the plurality of images and generates, based on the best-match image, an updated current pose estimate of the aircraft. The apparatus outputs the updated current pose estimate to a flight control computer of the aircraft, for use in guidance of the aircraft on a final approach.