Autonomous Navigation of Unmanned Aerial Vehicle: Investigating Architectures and Techniques for a Flexible Platform

Abstract

Unmanned aerial vehicles (UAVs) have a significant impact on the aircraft industry and are used for both military and civilian purposes. The partial or full automation of UAVs is an active research domain that is attracting interest from both the industrial and academic sectors. This paper presents a comprehensive review of UAVs, their architectures, and techniques used for autonomous navigation. Autonomous navigation requires the ability to recognize and avoid obstacles, which can be achieved using visual, non-visual, or hybrid techniques. These methods use computer vision and machine learning algorithms and are implemented on a computer system, which may or may not be embedded in the vehicle. After investigating the existing architectures and techniques, a flexible platform is designed. This platform combines the advantages of the existing architectures, materials, and methods. It is based on a compact architecture with the possibility of external communication. This platform is mainly composed of a companion computer for complex tasks like obstacle detection and path planning during mission. When implementing this platform, it is crucial to take into account the capabilities of the companion computer processor and appropriately select techniques. Finally, the paper suggests potential areas for future research in this field.