Vision-Based UAV Formation Control

Keywords

Unmanned Aerial Vehicle
vision-based
Histogram of Oriented Gradients
cascade classifier
Leader-Follower
flight controller

Abstract

Conventional Unmanned Aerial Vehicle (UAV) formation requires radio communication between the UAVs and the ground control station. However, noise and interruption in the communication links may result in severe consequences. In this paper, vision-based line formation is introduced to eliminate the problems mentioned earlier. Individual UAV with a camera on-board can detect its neighboring UAVs by implementing the concept of the Histogram of Oriented Gradients (HOG) descriptors and cascade classifiers. Image processing such as removal of shadow is applied to improve the detection rates.

Upon successful detection, the Leader-Follower formation theory is applied to allow the follower UAV to distinguish and follow the path of the leader UAV. Lastly, to control the movement of the UAVs, a flight controller is designed and implemented on Simulink, based on the line formation requirements.