Unmanned aerial vehicles (UAVs) have found widespread applications in various fields and scenarios, prompting extensive research on their formation control and route planning. However, in extreme application scenarios such as complex electromagnetic battlefields, existing studies often rely on radio frequency communication for data exchange between UAVs or utilize external sensors like GPS for positioning, which are susceptible to interference. To address this, we propose a method for distributed formation control of UAVs that solely relies on visual and inertial data, without data exchange between UAVs or the use of external positioning sensors like GPS. Drawing inspiration from the flight formations of migratory birds, our approach employs cameras to detect the orientations of neighboring UAVs. A simulation environment is constructed using a game engine to train a YOLO object detection network for reliable detection of other UAVs. Sub-sequently, utilizing geometric relationships, the positions of UAVs in the formation coordinate system are calculated, and our proposed formation control algorithm is applied to adjust the UAVs to their desired formation positions. Moreover, we employ the SIFT object detection algorithm to match images from the downward-facing camera of the UAV to satellite maps for obtaining global coordinates. Our introduced randomized navigator election algorithm selects a leading UAV for guiding the formation with higher accuracy along the intended route. Through simulation experiments, we demonstrate that the proposed approach achieves comparable accuracy to traditional methods in most scenarios, while exhibiting strong robustness.
A Vision-Only Drone Formation Waypoint Navigation Method
2024-11-15
3454759 byte
Conference paper
Electronic Resource
English
Forum - Waypoint Navigation for Rivercraft?
Online Contents | 1999
|Waypoint Navigation Using Exclusion Zones
NTIS | 1995
|Waypoint Navigation Using Exclusion Zones
NTIS | 1995
|Area navigation waypoint designation standards
Tema Archive | 1974
|