A robust detect and avoid system for autonomous drone navigation
NexusTech · 10.03.2026 · 1 (2) · 2026004
DOI: 10.31893/tech.2026004 ↗Kurz gefasst
Fachartikel in der Zeitschrift NexusTech, online erschienen am 10.03.2026. Verfasst von N. Vishnu Venkatesh und 2 weiteren Autor:innen. Die Arbeit nennt 12 Literaturangaben.
Abstract (Originalsprache)
Autonomous drone swarms require reliable collision avoidance to operate safely in congested airspace, yet existing detect-and-avoid (DAA) systems rely on expensive hardware unsuitable for multi-agent deployment. This study presents a lightweight, embedded DAA framework integrating computer vision and artificial intelligence on a Raspberry Pi-Pixhawk platform. The system employs YOLOv8 for real-time object detection, Kalman filtering for tracking, and potential field theory for collision prediction and avoidance trajectory generation. Unlike existing systems requiring high-end processors, our approach achieves real-time performance on resource-constrained hardware through algorithmic optimization. Field trials across 120 test scenarios demonstrated a 96.7% collision-free rate with mean response latency of 74.1 ms, representing 1.8× improvement over baseline OpenCV systems. Statistical analysis confirmed significant performance advantages (p < 0.01) across varying environmental conditions. This framework establishes a cost-effective foundation for autonomous swarm operations applicable to surveillance, rescue, and tactical missions.
Zitieren
APA
Venkatesh, N. V., Das, P., & Rajiv, A. (2026). A robust detect and avoid system for autonomous drone navigation. NexusTech, 1(2), 2026004. https://doi.org/10.31893/tech.2026004
Vancouver
Venkatesh NV, Das P, Rajiv A. A robust detect and avoid system for autonomous drone navigation. NexusTech. 2026;1(2):2026004. doi:10.31893/tech.2026004
BibTeX
@article{Venkatesh2026,
author = {Venkatesh, N. Vishnu and Das, Priyanksha and Rajiv, Asha},
title = {A robust detect and avoid system for autonomous drone navigation},
journal = {NexusTech},
year = {2026},
volume = {1},
number = {2},
pages = {2026004},
doi = {10.31893/tech.2026004}
}
EUVIMED