Task Allocation of Multiple Unmanned Aerial Vehicles Based on Deep Transfer Reinforcement Learning

Yin, Yongfeng; Guo, Yang; Su, Qingran; Wang, Zhetao

doi:10.3390/drones6080215

Cited by 37 publications

(8 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It can be seen from Figure 1B that as the short-circuit pin is loaded, the resonant frequency f 1/2,2 of the TM 1/2,2 mode will be far away from f 3/2,0 and cannot be directly used to increase the bandwidth (Chen et al, 2022a;Yin et al, 2022;Jiang and Xu, 2023). The resonant frequency ( f mn ) of TM mn in traditional PIFA can be obtained from the cavity model theory:…”

Section: Implementation and Analysis Of Four-mode Microstrip Antennamentioning

confidence: 99%

Design of novel microstrip patch antenna for millimeter-wave B5G communications

Tiang,

Alsekait,

Khan

et al. 2024

Front. Mater.

View full text Add to dashboard Cite

Introduction: The simplicity of integration and co-type features of microstrip antennas make them intriguing for a broad variety of applications, particularly with the growing usage of mmWave bands in wireless communications and the constant rise in data transfer in communication situations.Method: This paper proposes a novel design of micrstrip patch antenna for mmWave B5G communication. The main idea is to realize four-mode antenna the operates in four different frequencies. The geometry is rectangular patch whose resonance frequency is adjusted by varying the walls and pins of the structure.Results: Simulation results show that the proposed antenna design has improved fractional bandwidth and performance as compared with existing antennas.Discussion: The observed curve indicates that, in agreement with the modeling findings, there are four resonance spots in the operational frequency region of 2.5–3.4 GHz: 2.68 GHz, 2.9 GHz, 3.05 GHz, and 3.3 GHz, which correspond to TM1/2,0, TM3/2,0, and TMRS, respectively, and TM1/2,2 four resonant modes, within the frequency range, the observed antenna gain peak is around 9 dBi, which is consistent with the measured results.

show abstract

Section: Implementation and Analysis Of Four-mode Microstrip Antennamentioning

confidence: 99%

Design of novel microstrip patch antenna for millimeter-wave B5G communications

Tiang,

Alsekait,

Khan

et al. 2024

Front. Mater.

View full text Add to dashboard Cite

show abstract

“…Exploiting the ability to learn a non-linear function of such a technique with the reduced amount of parameters brings the skill of resolving the scalability issues together. The collaborative systems also need resource allocation and task assignment that leverage the strengths of each member to achieve optimal performance and results [26,27].…”

Section: Significance Of Scalabilitymentioning

confidence: 99%

“…The method is inspired by the multi-fidelity optimization [56]. The composed model is evaluated as Q lo and the model that takes into account the deep correction network is considered as a part of Q hi , as given in Equation (26). The deep correction network is the additive part that is nothing but another DQN.…”

Section: Deep Correctionmentioning

confidence: 99%

Genetic-Algorithm-Aided Deep Reinforcement Learning for Multi-Agent Drone Delivery

Tarhan,

Ure

2024

Drones

View full text Add to dashboard Cite

The popularity of commercial unmanned aerial vehicles has drawn great attention from the e-commerce industry due to their suitability for last-mile delivery. However, the organization of multiple aerial vehicles efficiently for delivery within limitations and uncertainties is still a problem. The main challenge of planning is scalability, since the planning space grows exponentially to the number of agents, and it is not efficient to let human-level supervisors structure the problem for large-scale settings. Algorithms based on Deep Q-Networks had unprecedented success in solving decision-making problems. Extension of these algorithms to multi-agent problems is limited due to scalability issues. This work proposes an approach that improves the performance of Deep Q-Networks on multi-agent delivery by drone problems by utilizing state decompositions for lowering the problem complexity, Curriculum Learning for handling the exploration complexity, and Genetic Algorithms for searching efficient packet-drone matching across the combinatorial solution space. The performance of the proposed method is shown in a multi-agent delivery by drone problem that has 10 agents and ≈1077 state–action pairs. Comparative simulation results are provided to demonstrate the merit of the proposed method. The proposed Genetic-Algorithm-aided multi-agent DRL outperformed the rest in terms of scalability and convergent behavior.

show abstract

“…In recent years, deep RL (DRL), which combines RL and deep learning (DL), has emerged as an important research area in UAV control and decision-making. DRL alleviates the dimension explosion problem that easily occurs in traditional RL and has made great breakthroughs in areas such as robot control [21][22][23], scheduling optimization [24,25], and multi-agent collaboration [26][27][28][29]. Ma [30] proposed a task-assignment algorithm based on Deep Q Network (DQN) to support UAV swarm operations, which significantly improves the success rate of UAV swarm combat.…”

Section: Related Workmentioning

confidence: 99%

Task Assignment of UAV Swarms Based on Deep Reinforcement Learning

Liu

Wang

et al. 2023

Drones

View full text Add to dashboard Cite

UAV swarm applications are critical for the future, and their mission-planning and decision-making capabilities have a direct impact on their performance. However, creating a dynamic and scalable assignment algorithm that can be applied to various groups and tasks is a significant challenge. To address this issue, we propose the Extensible Multi-Agent Deep Deterministic Policy Gradient (Ex-MADDPG) algorithm, which builds on the MADDPG framework. The Ex-MADDPG algorithm improves the robustness and scalability of the assignment algorithm by incorporating local communication, mean simulation observation, a synchronous parameter-training mechanism, and a scalable multiple-decision mechanism. Our approach has been validated for effectiveness and scalability through both simulation experiments in the Multi-Agent Particle Environment (MPE) and a real-world experiment. Overall, our results demonstrate that the Ex-MADDPG algorithm is effective in handling various groups and tasks and can scale well as the swarm size increases. Therefore, our algorithm holds great promise for mission planning and decision-making in UAV swarm applications.

show abstract

Task Allocation of Multiple Unmanned Aerial Vehicles Based on Deep Transfer Reinforcement Learning

Cited by 37 publications

References 10 publications

Design of novel microstrip patch antenna for millimeter-wave B5G communications

Design of novel microstrip patch antenna for millimeter-wave B5G communications

Genetic-Algorithm-Aided Deep Reinforcement Learning for Multi-Agent Drone Delivery

Task Assignment of UAV Swarms Based on Deep Reinforcement Learning

Contact Info

Product

Resources

About