UAV Vision Tracking Control for Obstacles and Poor GPS
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current aerial vehicle tracking systems lack intuitive and easy-to-use control methods, requiring manual piloting skills and limited real-time automatic control, especially in environments with obstacles or poor GPS signal quality, making them ineffective for tracking groups of moving objects or objects without defined features.
Innovation Solution
A system that uses a human-system interface to control unmanned aerial vehicles (UAVs) with processors configured to select modes for movement based on user input, allowing automatic detection and tracking of targets, and generating paths to navigate around obstacles, even in environments with poor GPS signal quality or complex object formations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual piloting control is used to operate the aerial vehicle, then the operator can control the vehicle to track targets and avoid obstacles, but the operator requires aviation experience and manual skill, reducing ease of operation
Solution Approach 1:
The aerial vehicle performs self-tracking of the target object using its onboard imaging device and processing unit. The vehicle automatically detects the target, calculates spatial relationships, and adjusts its flight path without requiring manual piloting, thereby improving ease of operation while maintaining system simplicity
Solution Approach 2:
The patent replaces manual mechanical piloting control with an automated vision-based tracking system. The processing unit substitutes for the human operator's manual control functions by automatically processing images, detecting targets, and generating flight control commands, thus improving ease of operation without significantly increasing device complexity
2Extent of automation
If vision-based tracking methods are used to track the target object, then the aerial vehicle can automatically detect and track the target, but the method is inadequate when obstacles appear in the flight path
Solution Approach 1:
The patent segments the tracking and navigation functions into distinct modules: target detection using the imaging device, spatial relationship calculation by the processing unit, and flight path planning that separately considers obstacles. This segmentation allows the system to maintain automated target tracking while independently handling obstacle avoidance, thereby improving reliability without reducing automation extent
Solution Approach 2:
The patent extends the tracking system from two-dimensional image processing to three-dimensional spatial reasoning. The processing unit calculates three-dimensional spatial relationships between the aerial vehicle, target object, and obstacles, enabling the system to navigate around obstacles while maintaining automated target tracking, thus improving reliability while preserving extent of automation
3Adaptability or versatility
If GPS-based tracking methods are used, then the aerial vehicle can navigate based on location data, but the method is limited by GPS signal quality and cannot track objects without GPS apparatus
Solution Approach 1:
The patent replaces GPS-based navigation with vision-based relative positioning. The processing unit calculates the aerial vehicle's position and orientation relative to the target object using image data from the onboard imaging device, eliminating dependence on GPS signals and external GPS apparatus on the target object. This substitution improves adaptability to environments with poor GPS signal quality while maintaining reliable tracking
Solution Approach 2:
The imaging device serves multiple functions: it acts as both the target detection device and the positioning reference. By using the same imaging device for both target acquisition and self-positioning relative to the target, the system achieves versatility in tracking objects without GPS apparatus while maintaining reliable navigation independent of GPS signal quality
Data Source
AI summary
A method for controlling an unmanned aerial vehicle (UAV) includes receiving, by a processor of the UAV, a plurality of images captured by an imaging device coupled to the UAV, identifying, by the processor, a target in at least one image of the plurality of images, determining, by the processor, whether the target is a stationary target or a moving target based on analyzing the plurality of images, and automatically effecting, by the processor, movement of the UAV based on determining the target is the stationary target or the moving target.