Zobrazit v EDS

Wildfire Front Monitoring With Multiple UAVs Using Deep Q-Learning

Uloženo v:
Podrobná bibliografie
Název: Wildfire Front Monitoring With Multiple UAVs Using Deep Q-Learning
Autoři: Alberto Viseras, Michael Meissner, Juan Marchal
Zdroj: IEEE Access, Vol 13, Pp 123269-123281 (2025)
Informace o vydavateli: Institute of Electrical and Electronics Engineers (IEEE), 2025.
Rok vydání: 2025
Témata: 0209 industrial biotechnology, multi-robot systems, intelligent robots, 02 engineering and technology, Intelligent robots, Unmanned aerial vehicles, TK1-9971, mobile robots, 13. Climate action, Multi-robot systems, Mobile robots, 0202 electrical engineering, electronic engineering, information engineering, Electrical engineering. Electronics. Nuclear engineering, unmanned aerial vehicles, Robot learning, Robot control, robot control
Popis: Wildfires destroy thousands of hectares every summer all over the globe. To provide an effective response and to mitigate wildfires impact, firefighters require a real-time monitoring of the fire front. This article proposes a cooperative reinforcement learning (RL) framework that allows a team of autonomous unmanned aerial vehicles (UAVs) to learn how to monitor a fire front. In the literature, independent Q-learners were proposed to solve a wildfire monitoring task with two UAVs. Here we propose a framework that can be easily extended to a larger number of UAVs. Our framework builds on two methods: multiple single trained Q-learning agents (MSTA) and value decomposition networks (VDN). MSTA trains a single UAV controller, which is then “copied” to each of the UAVs in the team. In contrast, VDN trains agents to learn how to cooperate. We benchmarked in simulations our two considered methods – MSTA and VDN – against two state-of-the-art approaches: independent Q-learners and a joint Q-learner. Simulation results show that our considered methods outperform state-of-the-art approaches in a wildfire front monitoring task with up to 9 fixed-wing and multi-copter UAVs.
Druh dokumentu: Article
Conference object
ISSN: 2169-3536
DOI: 10.1109/access.2021.3055651
Přístupová URL adresa: https://ieeexplore.ieee.org/ielx7/6287639/6514899/09340340.pdf
https://doaj.org/article/da398fa0b1ba4758b3db178df5304204
https://ieeexplore.ieee.org/abstract/document/9340340
https://elib.dlr.de/140953/
Rights: CC BY
CC BY NC ND
Přístupové číslo: edsair.doi.dedup.....76ed5b4fe8f085e6c5ae6d663028877f
Databáze: OpenAIRE
Popis
Abstrakt:Wildfires destroy thousands of hectares every summer all over the globe. To provide an effective response and to mitigate wildfires impact, firefighters require a real-time monitoring of the fire front. This article proposes a cooperative reinforcement learning (RL) framework that allows a team of autonomous unmanned aerial vehicles (UAVs) to learn how to monitor a fire front. In the literature, independent Q-learners were proposed to solve a wildfire monitoring task with two UAVs. Here we propose a framework that can be easily extended to a larger number of UAVs. Our framework builds on two methods: multiple single trained Q-learning agents (MSTA) and value decomposition networks (VDN). MSTA trains a single UAV controller, which is then “copied” to each of the UAVs in the team. In contrast, VDN trains agents to learn how to cooperate. We benchmarked in simulations our two considered methods – MSTA and VDN – against two state-of-the-art approaches: independent Q-learners and a joint Q-learner. Simulation results show that our considered methods outperform state-of-the-art approaches in a wildfire front monitoring task with up to 9 fixed-wing and multi-copter UAVs.
ISSN:21693536
DOI:10.1109/access.2021.3055651