Spatially Adaptive Positioning for Molecular Geometry Inspired Aerial Networks

Abstract

The advances in unmanned aerial vehicle (UAV) systems in terms of autonomy, reliability, communication and cost efficiency have lead to the employment of flying ad hoc networks (FANETs) with multiple UAVs in various applications such as surveillance, search and rescue missions, or aerial monitoring. The topology control and node positioning are challenging for FANET applications, which require the analysis of three dimensional space. Molecular geometry based approaches have been proposed for FANETs to overcome these challenges. In this paper, we propose a spatially adaptive positioning approach for FANETs. We define the collision probabilities for UAV positions and use them with boundary planes among UAVs in three dimensional space to analyze the molecular geometry based approaches and to show their efficiency. Then this analysis is used for the dynamic adaptation of the geometries for spaces with spatial constraints. We present a series of simulation scenarios to show the performance of our approach in coverage, collision probability, convergence time and cost in challenging environments.