Machine Learning for Autonomous Navigation and Collision Avoidance in UAVs

• Published in: Proceedings (International Confernce on Computational Intelligence and Communication Networks) pp. 381 - 388
• Main Authors: Mon, Bisni Fahad, Hayajneh, Mohammad, Ali, Najah Abu, Ullah, Farman, Al Warafy, Abdulmalik, Saeed, Nasir
• Format: Conference Proceeding
• Language: English
• Published: IEEE 22.12.2024
• Subjects: Adaptation models; Autonomous aerial vehicles; Autonomous robots; Collision avoidance; communication; machine learning; Navigation; Path planning; Real-time systems; Sensors; Testing; UAV; Vehicle dynamics
• ISSN: 2472-7555
• DOI: 10.1109/CICN63059.2024.10847476

Machine learning techniques are revolutionizing the field of autonomous navigation and collision avoidance in Unmanned Aerial Vehicles (UAVs). The advancement of UAVs toward autonomous navigation methods is aided by the sensors they carry, which can gather vast amounts of data, including images. This data can be used to train using vision-based deep learning autonomous navigation techniques. This study explores machine learning techniques to tackle complex navigation tasks such as path planning, localization, mapping, and obstacle detection. It also highlights the challenges of implementing machine learning in real-time environments, focusing on data management, computational efficiency, and the adaptability of models to dynamic conditions. By addressing these factors, the paper offers a detailed overview of how machine learning can improve UAV performance and suggests future research directions in this rapidly evolving field.