Distributed deep reinforcement learning for autonomous aerial eVTOL mobility in drone taxi applications

Yun, Won Joon; Jung, Soyi; Kim, Joongheon; Kim, Jae Hyun

doi:10.1016/j.icte.2021.01.005

Cited by 48 publications

(15 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For traffic flow management with self-driving EVs, Qu et al [76] aim to reduce stop-and-go traffic waves. In an autonomous flying taxi system, Yun et al [147] penalize the RL system for any in-flight collisions.…”

Section: Management Of User Discomfortmentioning

confidence: 99%

See 1 more Smart Citation

Exploiting Battery Storages With Reinforcement Learning: A Review for Energy Professionals

2022

View full text Add to dashboard Cite

The transition to renewable production and smart grids is driving a massive investment to battery storages, and reinforcement learning (RL) has recently emerged as a potentially disruptive technology for their control and optimization of battery storage systems. A surge of papers has appeared in the last two years applying reinforcement learning to the optimization of battery storages in buildings, energy communities, energy harvesting Internet of Things networks, renewable generation, microgrids, electric vehicles and plug-in hybrid electric vehicles. This article reviews these applications through 4 different perspectives. Firstly, the type of optimization problem is analyzed; the literature can be divided to approaches that optimize either financial targets or energy efficiency. Secondly, the approaches for handling user comfort are analyzed for applications that may impact a human user. Thirdly, this paper discusses the approach to model and reduce battery degradation. Fourthly, the articles are categorized by application context and applications likely to attract a high amount of research are identified. The paper concludes with a list of unresolved challenges.

show abstract

Section: Management Of User Discomfortmentioning

confidence: 99%

“…ii) EV FLEET For fleets of self-driving EVs, minimization targets include charging costs [124], charging times [120], time spent not carrying passengers [110] and battery exhaustion [147].…”

Section: B) Charging I) Carmentioning

confidence: 99%

Exploiting Battery Storages With Reinforcement Learning: A Review for Energy Professionals

2022

View full text Add to dashboard Cite

show abstract

“…, μ N } is the set of target policies, D is the replay memory, and θ i− Q is the parameter of the target Q network for agent i. To update each agent's actor parameters, the deterministic policy gradient is used to maximize J(θ i μ ) as shown in the equation (13).…”

Section: Multi-agent Ddpgmentioning

confidence: 99%

“…Many researches have been done on hunting with multi-robot systems problems. Among them, there are many methods for hunting a target by means of several mobile robots that are based on generative adversarial network [1], dynamic prediction [2] and Deep Reinforcement Learning (DRL) [11,13]. The nature inspired methods [4] are effective in chasing a dynamic target with random behavior in real time in unexpected environments.…”

Section: Introductionmentioning

confidence: 99%

Hybrid Formation Control for Multi-Robot Hunters Based on Multi-Agent Deep Deterministic Policy Gradient

Hamed

Hamlich

2021

mendel

View full text Add to dashboard Cite

The cooperation between mobile robots is one of the most important topics of interest to researchers, especially in the many areas in which it can be applied. Hunting a moving target with random behavior is an application that requires robust cooperation between several robots in the multi-robot system. This paper proposed a hybrid formation control for hunting a dynamic target which is based on wolves’ hunting behavior in order to search and capture the prey quickly and avoid its escape and Multi Agent Deep Deterministic Policy Gradient (MADDPG) to plan an optimal accessible path to the desired position. The validity and the effectiveness of the proposed formation control are demonstrated with simulation results.

show abstract

“…In the study [9] in which the communication problem for swarm drones is examined, a distributed deep reinforcement learning approach is presented to control more than one UAV. In the study, which presents a taxi model for autonomously moving aircraft to reach its customers as soon as possible, [10] eVTOL-based drone taxi method has been developed using a deep reinforcement learning approach. The parameter sharing method is suggested in the training for swarm robots by testing the PPO algorithm in different environments [11].…”

Section: Introductionmentioning

confidence: 99%

Drone Tracking with Drone using Deep Learning

Tan

Karaköse

ÖZET

2022

IJCIT

View full text Add to dashboard Cite

With the development of technology, studies in fields such as artificial intelligence, computer vision and deep learning are increasing day by day. In line with these developments, object tracking and object detection studies have spread over wide areas. In this article, a study is presented by simulating two different drones, a leader and a follower drone, accompanied by deep learning algorithms. Within the scope of this study, it is aimed to perform a drone tracking with drone in an autonomous way. Two different approaches are developed and tested in the simulator environment within the scope of drone tracking. The first of these approaches is to enable the leader drone to detect the target drone by using object-tracking algorithms. YOLOv5 deep learning algorithm is preferred for object detection. A data set of approximately 2500 images was created for training the YOLOv5 algorithm. The Yolov5 object detection algorithm, which was trained with the created data set, reached a success rate of approximately 93% as a result of the training. As the second approach, the object-tracking algorithm we developed is used. Trainings were carried out in the simulator created in the Matlab environment. The results are presented in detail in the following sections. In this article, some artificial neural networks and some object tracking methods used in the literature are explained.

show abstract

Distributed deep reinforcement learning for autonomous aerial eVTOL mobility in drone taxi applications

Cited by 48 publications

References 8 publications

Exploiting Battery Storages With Reinforcement Learning: A Review for Energy Professionals

Exploiting Battery Storages With Reinforcement Learning: A Review for Energy Professionals

Hybrid Formation Control for Multi-Robot Hunters Based on Multi-Agent Deep Deterministic Policy Gradient

Drone Tracking with Drone using Deep Learning

Contact Info

Product

Resources

About