An Artificial Intelligence-Enhanced UAV Configuration for Object Detection, Instant Communication, and Real-Time Information Transmission
2nd International Conference on Science, Engineering Management and Information Technology, SEMIT 2023, Ankara, Turkey, 14 - 15 September 2023, vol.2199 CCIS, pp.136-147, (Full Text)
- Publication Type: Conference Paper / Full Text
- Volume: 2199 CCIS
- Doi Number: 10.1007/978-3-031-72287-5_8
- City: Ankara
- Country: Turkey
- Page Numbers: pp.136-147
- Keywords: Communication, Data Transfer, Swarm, UAV
- Erciyes University Affiliated: Yes
Abstract
This paper presents a UAV design tailored for target detection in both individual and swarm surveillance flight missions. The primary focus is to enhance collaborative efforts within a UAV swarm for achieving common objectives. The design incorporates a communication mechanism optimized for swarm dynamics, equipped with hardware for efficient coordination. To facilitate seamless communication with other UAVs and the ground control station, a robust network architecture addresses interconnection and bandwidth requirements. Flight control is achieved through the utilization of a PixHawk 4 flight controller, a Raspberry Pi 4 for general-purpose computing, and an NVIDIA Jetson Nano for real-time object detection. The study addresses the following key aspects: optimizing communication within the UAV swarm, overcoming limitations in QGroundControl user interface for improved performance, and enhancing compatibility between QGroundControl software and detection applications. Additionally, future research opportunities, such as implementing video stream authorization over RTSP and supporting multi-UAV mission planning considering mesh network limitations, were identified. The findings highlight the potential for further improvements in UAV swarm operations, specifically in addressing communication challenges and refining mission planning interfaces. By considering these aspects, the work aims to contribute to the advancement of UAV technologies for effective surveillance and target detection in real-world applications.