Drone Target Tracking With GPS-Visual Fusion Framing Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Rotary-wing drones face challenges in accurately tracking and framing moving targets due to the inaccuracies and low update rates of GPS coordinates, leading to poor framing or the target being out of frame.
Innovation Solution
A system that combines GPS data with image analysis to adjust the camera's viewing angle and generate flight instructions using a hybrid coordinate extended Kalman filter, merging geographic and angular position information to improve tracking accuracy, and utilizing image analysis at a faster rate than GPS updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If GPS coordinates are used to track the target position, then the drone can autonomously follow the target, but the framing accuracy deteriorates due to GPS inaccuracies and low update rates
Solution Approach 1:
The patent merges GPS-based autonomous tracking with image analysis-based target detection. The system combines data from both sources to determine the target position, allowing the drone to maintain autonomous operation while improving positioning accuracy through visual feedback from the camera.
Solution Approach 2:
The patent introduces image analysis as an intermediary system that bridges the gap between GPS coordinates and actual target position. The image analysis provides real-time visual verification and correction of GPS-based positioning, acting as a mediator that reconciles the imprecise GPS data with the actual target location.
2Speed
If GPS update rate is increased to improve tracking responsiveness, then the target can be tracked more dynamically, but the system complexity and computational load increase
Solution Approach 1:
The patent implements dynamic tracking by continuously analyzing images at the camera frame rate and adjusting the drone's flight path in real-time. The system dynamically updates the target position based on visual detection, allowing responsive tracking without requiring increased GPS update rates, thus avoiding the complexity associated with high-frequency GPS updates.
3Device complexity
If only GPS data is used for target tracking, then the system remains simple, but the framing accuracy and target positioning deteriorate
Solution Approach 1:
The patent replaces the mechanical GPS positioning system with an optical-based image analysis system for target detection. By using computer vision algorithms to identify and track the target in camera images, the system achieves superior positioning accuracy without significantly increasing overall system complexity, as the camera and processing are already integral to the drone.
Data Source
Figure 1~3
Figure 4
AI summary
The movements of the drone (D) are defined by piloting commands to shoot a target carrying the ground station (T). The system comprises means for adjusting the viewing angle of the camera (C) during the movements of the drone and of the target, so that the images are framed on the target, and means for generating flight instructions in such a way that the distance between drone and target responds to determined rules, these means being based on a determination of the GPS geographical position of the target with respect to the GPS geographical position of the drone, and of the angular position of the target with respect to the main axis of the drone. These means are also based on the analysis of a non-geographical signal produced by the target and received by the drone. The system makes it possible to overcome the uncertainty of the GPS systems fitted to this type of device.