Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10155944
Zhanibek Darush, Mikhail Martynov, A. Fedoseev, A. Shcherbak, D. Tsetserukou
The continuous monitoring by drone swarms remains a challenging problem due to the lack of power supply and the inability of drones to land on uneven surfaces. Heterogeneous swarms can support longer inspections, however, their capabilities are limited by the mobility of wheeled and legged robots in a cluttered environment.In this paper, we propose a novel concept SwarmGear for autonomous inspection. It leverages a heterogeneous swarm of drones that investigates the environment in a leader-follower formation. The leader drone is able to land on rough terrain and traverse it by four compliant robotic pedipulators, possessing both the functionalities of an aerial and mobile robot. To preserve the formation of the swarm during its motion, virtual impedance links were developed between the leader and the follower drones.The experiments revealed low crosstrack error (mean value is of 2.2 cm and max one is of 5.3 cm with the Type 2 gait) and the ability of the leader drone to move with a 190 mm step length. Four types of drone formation were considered. The best formation was applied for experiments and showed low overall crosstrack error for the swarm (mean 3.9 cm for the Type 1 gait and 3.3 cm for the Type 2 gait).
由于缺乏电力供应和无人机无法在凹凸不平的地面上着陆，无人机群的持续监测仍然是一个具有挑战性的问题。异构蜂群可以支持更长时间的检查，然而，它们的能力受到轮式和腿式机器人在混乱环境中的移动性的限制。在本文中，我们提出了一个新的概念swarm - gear用于自主检测。它利用一群异质的无人机，以领导-追随者的形式调查环境。领头的无人机能够降落在崎岖的地形上，并通过四个顺从的机器人爬行器穿越它，同时具有空中和移动机器人的功能。为了保证蜂群在运动过程中保持队形，在领头和跟随无人机之间建立了虚拟阻抗链路。实验结果表明，该系统具有较低的跨轨误差(2型步态的平均值为2.2 cm，最大误差为5.3 cm)，并且能够以190 mm的步长移动。考虑了四种类型的无人机编队。采用最佳队形进行实验，结果表明，该队形整体交叉误差较低(1型步态平均3.9 cm, 2型步态平均3.3 cm)。
{"title":"SwarmGear: Heterogeneous Swarm of Drones with Morphogenetic Leader Drone and Virtual Impedance Links for Multi-Agent Inspection","authors":"Zhanibek Darush, Mikhail Martynov, A. Fedoseev, A. Shcherbak, D. Tsetserukou","doi":"10.1109/ICUAS57906.2023.10155944","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10155944","url":null,"abstract":"The continuous monitoring by drone swarms remains a challenging problem due to the lack of power supply and the inability of drones to land on uneven surfaces. Heterogeneous swarms can support longer inspections, however, their capabilities are limited by the mobility of wheeled and legged robots in a cluttered environment.In this paper, we propose a novel concept SwarmGear for autonomous inspection. It leverages a heterogeneous swarm of drones that investigates the environment in a leader-follower formation. The leader drone is able to land on rough terrain and traverse it by four compliant robotic pedipulators, possessing both the functionalities of an aerial and mobile robot. To preserve the formation of the swarm during its motion, virtual impedance links were developed between the leader and the follower drones.The experiments revealed low crosstrack error (mean value is of 2.2 cm and max one is of 5.3 cm with the Type 2 gait) and the ability of the leader drone to move with a 190 mm step length. Four types of drone formation were considered. The best formation was applied for experiments and showed low overall crosstrack error for the swarm (mean 3.9 cm for the Type 1 gait and 3.3 cm for the Type 2 gait).","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127935882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156218
E. Aldao, Gabriel Fontenla-Carrera, L. M. González-deSantos, H. González-Jorge
The use of UAVs for the delivery of emergency medical equipment, such as Automated External Defibrillators (AEDs), has the potential to significantly improve response times in life-threatening situations. This time saving can be crucial for the patient’s life. However, the flight in urban areas presents a great challenge due to its complexity and regulatory limitations on the flight over densely populated areas. Therefore, in this work, a UAV route calculation algorithm that minimizes flight time taking into account these restrictions is proposed. Firstly, the horizontal flight profile is computed, generating trajectories avoiding densely populated areas. Then, considering the elevation of the terrain, the vertical profile as well as the aircraft performances are optimized using a direct Optimal Control method. A practical study case was developed to demonstrate the capabilities of the developed implementation.
{"title":"UAV Path Planning for the delivery of emergency medical supplies","authors":"E. Aldao, Gabriel Fontenla-Carrera, L. M. González-deSantos, H. González-Jorge","doi":"10.1109/ICUAS57906.2023.10156218","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156218","url":null,"abstract":"The use of UAVs for the delivery of emergency medical equipment, such as Automated External Defibrillators (AEDs), has the potential to significantly improve response times in life-threatening situations. This time saving can be crucial for the patient’s life. However, the flight in urban areas presents a great challenge due to its complexity and regulatory limitations on the flight over densely populated areas. Therefore, in this work, a UAV route calculation algorithm that minimizes flight time taking into account these restrictions is proposed. Firstly, the horizontal flight profile is computed, generating trajectories avoiding densely populated areas. Then, considering the elevation of the terrain, the vertical profile as well as the aircraft performances are optimized using a direct Optimal Control method. A practical study case was developed to demonstrate the capabilities of the developed implementation.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126450036","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156157
S. R. Nekoo, A. Ollero
This paper presents an equivalent dynamic model for vertical regulation control of a flapping-wing flying robot. The model is presented based on the data of a series of flight experiments for an available platform. The system shows oscillations in motion in all experiments with an approximate frequency between [3. 5, 4. 5](Hz), changing within a limited range. The behavior of the equivalent model represents a system with base excitation. The displacement transmissibility ratio (TR) is found for the model to investigate the oscillatory behavior in the system during the flight. Reduction of the oscillations through the transmissibility ratio will decrease the uncertainty in flight and consequently, that could increase the success rate of perching on a branch (now it has a 10 - 15(cm) uncertain periodic motion); perching needs precision on the last meter approaching phase. An analytical expression for TR is presented which is used for parameter selection, tuning, and selection of the flapping frequency, as the base excitation source. The study shows that the robot works in a proper zone of the frequency ratio, and also, the sensitivity of the TR is high concerning the changes in the stiffness constant.
{"title":"Equivalent Vertical Dynamics of Flapping-Wing Flying Robot in Regulation Control: Displacement Transmissibility Ratio","authors":"S. R. Nekoo, A. Ollero","doi":"10.1109/ICUAS57906.2023.10156157","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156157","url":null,"abstract":"This paper presents an equivalent dynamic model for vertical regulation control of a flapping-wing flying robot. The model is presented based on the data of a series of flight experiments for an available platform. The system shows oscillations in motion in all experiments with an approximate frequency between [3. 5, 4. 5](Hz), changing within a limited range. The behavior of the equivalent model represents a system with base excitation. The displacement transmissibility ratio (TR) is found for the model to investigate the oscillatory behavior in the system during the flight. Reduction of the oscillations through the transmissibility ratio will decrease the uncertainty in flight and consequently, that could increase the success rate of perching on a branch (now it has a 10 - 15(cm) uncertain periodic motion); perching needs precision on the last meter approaching phase. An analytical expression for TR is presented which is used for parameter selection, tuning, and selection of the flapping frequency, as the base excitation source. The study shows that the robot works in a proper zone of the frequency ratio, and also, the sensitivity of the TR is high concerning the changes in the stiffness constant.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"137 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127343721","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156489
Dong-Hee Lee, Wooryong Park, Junhak Yi, Woohyun Byun, Soobin Huh, Woochul Nam
It is difficult for unmanned aerial vehicles to chase another micro-aircraft (MA) due to the small size and its fast maneuverability. Thus, this study developed a fast and accurate visual tracker for real-time inference. Then, a quadcopter was controlled to chase a target MA by considering the result of the visual tracker. Specifically, the pitch, throttle, and yaw of the quadcopter were determined by the PD controller based on the position, and the size of the MA in the image. The newly developed visual tracker comprises an adaptive search region (SR) and a fully convolutional neural network. The size and the location of the SR were constantly adjusted over image frames by considering the tracking result of the MA in previous frames. Furthermore, if the size and the location of the SR are not precise enough, the SR was updated to minimize the tracking failure. Performance of the SR was improved by using the Kalman filter. In real flight experiments, the quadcopter equipped with the proposed model successfully chased the MA which randomly moved at approximately 5 m/s.
{"title":"Quadcopter Capable of Autonomously Chasing Micro-Aircraft with Real-Time Visual Tracker","authors":"Dong-Hee Lee, Wooryong Park, Junhak Yi, Woohyun Byun, Soobin Huh, Woochul Nam","doi":"10.1109/ICUAS57906.2023.10156489","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156489","url":null,"abstract":"It is difficult for unmanned aerial vehicles to chase another micro-aircraft (MA) due to the small size and its fast maneuverability. Thus, this study developed a fast and accurate visual tracker for real-time inference. Then, a quadcopter was controlled to chase a target MA by considering the result of the visual tracker. Specifically, the pitch, throttle, and yaw of the quadcopter were determined by the PD controller based on the position, and the size of the MA in the image. The newly developed visual tracker comprises an adaptive search region (SR) and a fully convolutional neural network. The size and the location of the SR were constantly adjusted over image frames by considering the tracking result of the MA in previous frames. Furthermore, if the size and the location of the SR are not precise enough, the SR was updated to minimize the tracking failure. Performance of the SR was improved by using the Kalman filter. In real flight experiments, the quadcopter equipped with the proposed model successfully chased the MA which randomly moved at approximately 5 m/s.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127647472","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156277
Deeshant Sharma, Annu, Naga Praveen Babu Mannam, P. Rajlakshmi
Nowadays, aerial and ground robots can be made cheaper and lighter which makes it possible to deploy them in large numbers and drone swarms have the potential to increase efficiency and safety in certain applications, as well as provide new capabilities that would not be possible with a single drone. Successful implementation of swarm cooperative applications requires low-latency communications and real-time localization. In this paper, we proposed a cloud-based control system architecture to dynamically control the drone swarm or UAV formation in the 3D space using the mobile application. A group of UAVs determines their location using an integrated ultra-wideband module. The base station is connected to the cloud platform (google firebase in our case) which is again connected to a mobile app to get the position and formation commands directly from the user using an interactive interface. The base station will get these commands and position control information from the cloud. The base station then sends the next setpoint to each UAV, enabling UAVs to form a real-time user-controlled formation and fly autonomously until the next command. Our experiment results show that the latency in this architecture is in a range of 0.8 to 1.41 sec. with fixed anchors, the localization error is less than 5 cm.
{"title":"Cloud-Based Control of Drone Swarm with Localization via Ultra-Wideband","authors":"Deeshant Sharma, Annu, Naga Praveen Babu Mannam, P. Rajlakshmi","doi":"10.1109/ICUAS57906.2023.10156277","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156277","url":null,"abstract":"Nowadays, aerial and ground robots can be made cheaper and lighter which makes it possible to deploy them in large numbers and drone swarms have the potential to increase efficiency and safety in certain applications, as well as provide new capabilities that would not be possible with a single drone. Successful implementation of swarm cooperative applications requires low-latency communications and real-time localization. In this paper, we proposed a cloud-based control system architecture to dynamically control the drone swarm or UAV formation in the 3D space using the mobile application. A group of UAVs determines their location using an integrated ultra-wideband module. The base station is connected to the cloud platform (google firebase in our case) which is again connected to a mobile app to get the position and formation commands directly from the user using an interactive interface. The base station will get these commands and position control information from the cloud. The base station then sends the next setpoint to each UAV, enabling UAVs to form a real-time user-controlled formation and fly autonomously until the next command. Our experiment results show that the latency in this architecture is in a range of 0.8 to 1.41 sec. with fixed anchors, the localization error is less than 5 cm.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114078319","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156349
Gulnihal Kucuksayacigil
In this work, a recursive algorithm has been developed for heterogeneous network distributed systems (NDS) communicating over an undirected network to solve H∞ optimal distributed tracking control problem of continuous-time systems as a convex problem. Recent studies on NDS have studied the tracking control problem with decentralized performance functions defined for each subsystem in the network, on the contrary, a global performance function has been defined in this work for the whole NDS. An optimal distributed control problem has been defined as a sequential convex optimization problem benefiting off-policy reinforcement learning with sparsity constraints introduced on the symmetric positive definite matrix. Finally, the efficacy of the proposed algorithm is shown on a group of heterogeneous unmanned aerial vehicles (UAV) communicating over an undirected network.
{"title":"H∞ Optimal Distributed Tracking Control Algorithm for Network Distributed Systems with an Application to UAV","authors":"Gulnihal Kucuksayacigil","doi":"10.1109/ICUAS57906.2023.10156349","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156349","url":null,"abstract":"In this work, a recursive algorithm has been developed for heterogeneous network distributed systems (NDS) communicating over an undirected network to solve H∞ optimal distributed tracking control problem of continuous-time systems as a convex problem. Recent studies on NDS have studied the tracking control problem with decentralized performance functions defined for each subsystem in the network, on the contrary, a global performance function has been defined in this work for the whole NDS. An optimal distributed control problem has been defined as a sequential convex optimization problem benefiting off-policy reinforcement learning with sparsity constraints introduced on the symmetric positive definite matrix. Finally, the efficacy of the proposed algorithm is shown on a group of heterogeneous unmanned aerial vehicles (UAV) communicating over an undirected network.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122169783","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156106
Simone Martini, M. Stefanovic, A. Rizzo, M. J. Rutherford, P. Livreri, K. Valavanis
A benchmark framework to test, evaluate, and compare different quadrotor controllers is presented. A detailed nonlinear quadrotor model and a set of six mission scenarios are used to evaluate seven state-of-the-art linear and nonlinear controllers. The quadrotor model is based on the Lagrange formulation and includes aerodynamic and gyroscopic effects, allows for sensor feedback noise to be introduced, and account for first order motor dynamics with input saturation. Simulated mission scenarios include realistic disturbances such as abrupt change of mass, wind gust, and aggressive flight maneuvers. The benchmark framework is the primary contribution of this research; the framework allows for performance comparison of multiple control architectures and implementations, and, the resulting open access testbed is made available to other researchers. Moreover, the same framework may be used to conduct simulated experiments (using ROS/Gazebo, X-Plane, or other software tools), and, with minor modifications, to compare controller performance based on real flights.
{"title":"A Benchmark Framework for Testing, Evaluation, and Comparison of Quadrotor Linear and Nonlinear Controllers","authors":"Simone Martini, M. Stefanovic, A. Rizzo, M. J. Rutherford, P. Livreri, K. Valavanis","doi":"10.1109/ICUAS57906.2023.10156106","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156106","url":null,"abstract":"A benchmark framework to test, evaluate, and compare different quadrotor controllers is presented. A detailed nonlinear quadrotor model and a set of six mission scenarios are used to evaluate seven state-of-the-art linear and nonlinear controllers. The quadrotor model is based on the Lagrange formulation and includes aerodynamic and gyroscopic effects, allows for sensor feedback noise to be introduced, and account for first order motor dynamics with input saturation. Simulated mission scenarios include realistic disturbances such as abrupt change of mass, wind gust, and aggressive flight maneuvers. The benchmark framework is the primary contribution of this research; the framework allows for performance comparison of multiple control architectures and implementations, and, the resulting open access testbed is made available to other researchers. Moreover, the same framework may be used to conduct simulated experiments (using ROS/Gazebo, X-Plane, or other software tools), and, with minor modifications, to compare controller performance based on real flights.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126088742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10155816
Ahmet Cagri Arican, Engin Hasan Çopur, G. Inalhan, M. U. Salamci
For linear systems, a state feedback control law can be easily designed to keep all closed-loop poles inside a specified disk since the locations of the poles are unique. However, its application to nonlinear systems is not so simple. Therefore, this paper introduces a new pole placement method, named as State Dependent Regional Pole Assignment, for nonlinear systems. This proposed method produces a state dependent feedback control law, enabling the eigenvalues of the closed-loop matrix to be placed in a specified disk to achieve the desired control performance characteristics. The effectiveness of the method is tested on the 3 DOF Helicopter experimental setup. To verify its effectiveness, the experimental results of the nonlinear method are compared with those of the linear method.
{"title":"State Dependent Regional Pole Assignment Controller Design for a 3-DOF Helicopter Model","authors":"Ahmet Cagri Arican, Engin Hasan Çopur, G. Inalhan, M. U. Salamci","doi":"10.1109/ICUAS57906.2023.10155816","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10155816","url":null,"abstract":"For linear systems, a state feedback control law can be easily designed to keep all closed-loop poles inside a specified disk since the locations of the poles are unique. However, its application to nonlinear systems is not so simple. Therefore, this paper introduces a new pole placement method, named as State Dependent Regional Pole Assignment, for nonlinear systems. This proposed method produces a state dependent feedback control law, enabling the eigenvalues of the closed-loop matrix to be placed in a specified disk to achieve the desired control performance characteristics. The effectiveness of the method is tested on the 3 DOF Helicopter experimental setup. To verify its effectiveness, the experimental results of the nonlinear method are compared with those of the linear method.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"97 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122903600","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156126
Mohamed Mokhtar, A. El-Badawy
A deep reinforcement learning-based control framework has been proposed in this paper to achieve autonomous navigation and control of a quadrotor. Cascaded reinforcement learning agents form the control framework. First, a path following (PF) agent controls the quadrotor’s tracking behavior by directly mapping environment states into motor commands. The second agent modifies the desired path to avoid any detected obstacles along the path. The obstacle avoidance (OA) agent achieves this task by adding an offset distance deflection to the tracking error before sending it to the path-following agent. Generalization of the obstacle avoidance behavior in three-dimensional space was achieved by the usage of frame transformation. The two agents were trained using the "Twin Delayed Deep Deterministic Policy Gradient" (TD3) algorithm, and the developed framework succeeded in avoiding multiple obstacles of different sizes and configurations in simulation.
为了实现四旋翼飞行器的自主导航和控制，提出了一种基于深度强化学习的控制框架。级联强化学习代理构成控制框架。首先，路径跟随(PF)代理控制四旋翼的跟踪行为直接映射到电机命令的环境状态。第二个智能体修改期望的路径以避免路径上检测到的任何障碍物。避障(OA)代理在将跟踪错误发送给路径跟踪代理之前，通过在跟踪错误上添加偏移距离偏转来完成此任务。利用帧变换实现了避障行为在三维空间的泛化。采用“双延迟深度确定性策略梯度”(Twin Delayed Deep Deterministic Policy Gradient, TD3)算法对两个智能体进行训练，开发的框架在仿真中成功地避开了多个不同大小和配置的障碍物。
{"title":"Autonomous Navigation and Control of a Quadrotor Using Deep Reinforcement Learning","authors":"Mohamed Mokhtar, A. El-Badawy","doi":"10.1109/ICUAS57906.2023.10156126","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156126","url":null,"abstract":"A deep reinforcement learning-based control framework has been proposed in this paper to achieve autonomous navigation and control of a quadrotor. Cascaded reinforcement learning agents form the control framework. First, a path following (PF) agent controls the quadrotor’s tracking behavior by directly mapping environment states into motor commands. The second agent modifies the desired path to avoid any detected obstacles along the path. The obstacle avoidance (OA) agent achieves this task by adding an offset distance deflection to the tracking error before sending it to the path-following agent. Generalization of the obstacle avoidance behavior in three-dimensional space was achieved by the usage of frame transformation. The two agents were trained using the \"Twin Delayed Deep Deterministic Policy Gradient\" (TD3) algorithm, and the developed framework succeeded in avoiding multiple obstacles of different sizes and configurations in simulation.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128020113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-06DOI: 10.1109/ICUAS57906.2023.10156565
Sara Roos-Hoefgeest, Jonathan Cacace, Vincenzo Scognamiglio, Ignacio Álvarez, R. C. González, Fabio Ruggiero, V. Lippiello
Inspecting and maintaining industrial plants is an important and emerging field in robotics. A particular case is represented by the inspection of oil and gas refinery facilities consisting of different long pipe racks to be inspected repeatedly. This task is costly in terms of human safety and operation costs due to the high altitude location in which the pipes are placed. In this domain, we propose a visual inspection system for unmanned aerial vehicles (UAVs), allowing the autonomous tracking and navigation of the center line of the industrial pipe. The proposed approach exploits a depth sensor to generate the control data for the aerial platform and, at the same time, highlight possible pipe defects. A set of simulated and real experiments in a GPS-denied environment have been carried out to validate the visual inspection system.
{"title":"A Vision-based Approach for Unmanned Aerial Vehicles to Track Industrial Pipes for Inspection Tasks","authors":"Sara Roos-Hoefgeest, Jonathan Cacace, Vincenzo Scognamiglio, Ignacio Álvarez, R. C. González, Fabio Ruggiero, V. Lippiello","doi":"10.1109/ICUAS57906.2023.10156565","DOIUrl":"https://doi.org/10.1109/ICUAS57906.2023.10156565","url":null,"abstract":"Inspecting and maintaining industrial plants is an important and emerging field in robotics. A particular case is represented by the inspection of oil and gas refinery facilities consisting of different long pipe racks to be inspected repeatedly. This task is costly in terms of human safety and operation costs due to the high altitude location in which the pipes are placed. In this domain, we propose a visual inspection system for unmanned aerial vehicles (UAVs), allowing the autonomous tracking and navigation of the center line of the industrial pipe. The proposed approach exploits a depth sensor to generate the control data for the aerial platform and, at the same time, highlight possible pipe defects. A set of simulated and real experiments in a GPS-denied environment have been carried out to validate the visual inspection system.","PeriodicalId":379073,"journal":{"name":"2023 International Conference on Unmanned Aircraft Systems (ICUAS)","volume":"275 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115878875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：