UAV Vision Navigation Using Optic Flow in GPS-Denied Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmanned aerial vehicles (UAVs) face challenges in navigating without global navigation information, such as GPS, which can lead to position drift and inaccurate location estimation over time.
Innovation Solution
A system and method utilizing an image sensor to determine optic flow due to vehicle motion and selecting a target for guidance, allowing the UAV to maneuver without relying on GPS by commanding the vehicle based on the difference in optic flow and ground speed estimates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If global navigation systems (GPS) are used for UAV navigation, then location accuracy is improved, but the system becomes vulnerable to signal loss and degradation over time
Solution Approach 1:
The patent introduces an image sensor as an intermediary tool to capture visual information of the environment and selected targets. This visual information serves as a mediator between the UAV and the ground, enabling the system to estimate position and maintain navigation accuracy without relying solely on GPS signals. The image sensor captures optic flow and target position data that act as intermediate measurements for navigation calculations.
Solution Approach 2:
The patent replaces the GPS-based electronic navigation system with a vision-based navigation approach using image sensors. Instead of relying on satellite signals and electronic position calculations, the system uses optical flow measurement and target tracking through image processing to determine UAV position and velocity, substituting the mechanical/electronic GPS system with an optical measurement system.
2Measurement precision
If inertial sensors are used for short-term position determination, then position accuracy is maintained initially, but estimates degrade rapidly over time
Solution Approach 1:
The patent implements continuous visual measurement through the image sensor to continuously update position estimates. By continuously capturing images and calculating optic flow and target position, the system maintains ongoing correction of the inertial navigation drift, ensuring continuous useful action for position determination throughout the entire flight duration without time-limited accuracy.
Solution Approach 2:
The system uses feedback from the image sensor to continuously monitor and correct position estimates. The visual information about the environment and selected targets provides feedback signals that are processed to generate correction terms for the inertial navigation system, creating a closed-loop feedback mechanism that prevents drift accumulation over time.
3Extent of automation
If visual target tracking is implemented without GPS, then navigation autonomy is improved, but system complexity increases
Solution Approach 1:
The patent extracts and utilizes specific visual features from the image data, namely optic flow and selected target position, to enable navigation. By focusing on these extracted visual cues rather than processing the entire image data set, the system achieves navigation autonomy while managing computational complexity. The extraction of relevant visual information simplifies the processing requirements compared to comprehensive visual analysis.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables UAVs to maintain accurate navigation and target tracking for extended periods without GPS data, using image-based guidance to maintain a stable orbit around a selected target, enhancing their operational capabilities in GPS-denied environments.
Implementation Method 1
determining a difference between an optic flow due only to motion of the vehicle and the selected target of the image
Data Source
AI summary
A system and method for navigating a vehicle comprising an image sensor in the absence of global positioning information is disclosed. In one embodiment, the method comprises accepting a user-selected target of an image produced by the imaging sensor, determining a difference between an optic flow due only to motion of the vehicle and the selected target of the image, determining a vehicle guidance command at least in part according to the difference between the optic flow of the selected target due to motion of the vehicle and the selected target of the image, and an estimate of a ground speed of the vehicle Vg, and commanding the vehicle at least in part according to the vehicle guidance command. Another embodiment is evidenced by an apparatus having a processor and a communicatively coupled memory storing processor instructions for performing the foregoing operations.