Please use this identifier to cite or link to this item: https://hdl.handle.net/11499/56725
Title: Modeling Energy Consumption of Small Drones for Swarm Missions
Authors: Cabuk, U.C.
Tosun, M.
Dagdeviren, O.
Ozturk, Yusuf
Keywords: Atmospheric modeling
Cost function
Data models
Drone
Drones
Energy consumption
energy model
machine learning
Payloads
regression
swarm
Topology
UAV
XGBoost
Cost functions
Decision making
Drones
Energy efficiency
Energy utilization
Latexes
Learning systems
Linear regression
Swarm intelligence
Topology
Atmospheric modeling
Cost-function
Energy model
Energy-consumption
Machine-learning
Payload
Regression
Swarm
Theoretical modeling
Xgboost
Antennas
Publisher: Institute of Electrical and Electronics Engineers Inc.
Abstract: Drones, particularly when deployed in swarms, hold immense potential for various applications, such as aerial imaging, delivery services, disaster response, and advanced surveillance. Their effective and efficient use, however, hinges on the accurate estimation of energy consumption. This study focuses on determining energy consumption patterns for small drones weighing less than 2 kg and with diameters under 1 m. Through an extensive series of test flights, empirical energy consumption data was collected and used to develop four distinct energy models: a theoretical model based on fundamental flight forces, a simple linear regression model, a cubic polynomial regression model, and a machine learning-based model using the XGBoost Regressor algorithm. A cost function for swarm topology control was also derived for assessing energy consumption during various activities, like connectivity restoration and formation change, facilitating more energy-efficient decision-making in swarm operations. Our findings highlighted an energy-efficient “valley” within the airspeed range, indicating that flying at speeds outside this range results in higher energy consumption. The machine learning model demonstrated superior accuracy within its training range, achieving an R<inline-formula> <tex-math notation="LaTeX">$^2$</tex-math> </inline-formula> of 0.9999, whereas the polynomial regression model was deemed best for extrapolation purposes, delivering an R<inline-formula> <tex-math notation="LaTeX">$^2$</tex-math> </inline-formula> of 0.966. Simple linear regression and theoretical models, although less accurate, can offer quick energy demand calculations and insights into the effects of hardware modifications, respectively. IEEE
URI: https://doi.org/10.1109/TITS.2024.3350042
https://hdl.handle.net/11499/56725
ISSN: 1524-9050
Appears in Collections:Mühendislik Fakültesi Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection
WoS İndeksli Yayınlar Koleksiyonu / WoS Indexed Publications Collection

Show full item record



CORE Recommender

WEB OF SCIENCETM
Citations

1
checked on Jul 1, 2024

Page view(s)

56
checked on May 27, 2024

Google ScholarTM

Check




Altmetric


Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.