Pub Date : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999695
Vinorth Varatharasan, Hyo-Sang Shin, A. Tsourdos, Nick Colosimo
Usually, Neural Networks models are trained with a large dataset of images in homogeneous backgrounds. The issue is that the performance of the network models trained could be significantly degraded in a complex and heterogeneous environment. To mitigate the issue, this paper develops a framework that permit to autonomously generate a training dataset in heterogeneous cluttered backgrounds. It is clear that the learning effectiveness of the proposed framework should be improved in complex and heterogeneous environments, compared with the ones with the typical dataset. In our framework, a state-of-the-art image segmentation technique called DeepLab is used to extract objects of interest from a picture and Chroma-key technique is then used to merge the extracted objects of interest into specific heterogeneous backgrounds. The performance of the proposed framework is investigated through empirical tests and compared with that of the model trained with the COCO dataset. The results show that the proposed framework outperforms the model compared. This implies that the learning effectiveness of the framework developed is superior to the models with the typical dataset.
{"title":"Improving Learning Effectiveness For Object Detection and Classification in Cluttered Backgrounds","authors":"Vinorth Varatharasan, Hyo-Sang Shin, A. Tsourdos, Nick Colosimo","doi":"10.1109/REDUAS47371.2019.8999695","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999695","url":null,"abstract":"Usually, Neural Networks models are trained with a large dataset of images in homogeneous backgrounds. The issue is that the performance of the network models trained could be significantly degraded in a complex and heterogeneous environment. To mitigate the issue, this paper develops a framework that permit to autonomously generate a training dataset in heterogeneous cluttered backgrounds. It is clear that the learning effectiveness of the proposed framework should be improved in complex and heterogeneous environments, compared with the ones with the typical dataset. In our framework, a state-of-the-art image segmentation technique called DeepLab is used to extract objects of interest from a picture and Chroma-key technique is then used to merge the extracted objects of interest into specific heterogeneous backgrounds. The performance of the proposed framework is investigated through empirical tests and compared with that of the model trained with the COCO dataset. The results show that the proposed framework outperforms the model compared. This implies that the learning effectiveness of the framework developed is superior to the models with the typical dataset.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125942927","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999671
A. Gàlffy, Florian Car, G. Schitter
This paper focuses on the integration and characterization of a 3D printed 5-hole probe for high-dynamic wind measurements for UAV. The probe is calibrated by means of a free jet wind tunnel and validated by dynamic angle of attack variations during an unmanned test flight. The angle of attack and sideslip angle measurements show an RMS error of 24 mrad with a peak-to-peak noise of 3mrad in the range of −20° to 20°. The airspeed measurement shows an RMS error of 0.19ms−1 with a peak-to-peak noise of 0.09ms−1 and a range up to 28ms−1. Therefore, the probe fulfills the requirements for the planned purpose of active turbulence suppression for fixed-wing UAV. Furthermore, the probe appears also suitable for a variety of other applications.
{"title":"Calibration and flight test of a 3D printed 5-hole probe for high-dynamic wind measurements for UAV","authors":"A. Gàlffy, Florian Car, G. Schitter","doi":"10.1109/REDUAS47371.2019.8999671","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999671","url":null,"abstract":"This paper focuses on the integration and characterization of a 3D printed 5-hole probe for high-dynamic wind measurements for UAV. The probe is calibrated by means of a free jet wind tunnel and validated by dynamic angle of attack variations during an unmanned test flight. The angle of attack and sideslip angle measurements show an RMS error of 24 mrad with a peak-to-peak noise of 3mrad in the range of −20° to 20°. The airspeed measurement shows an RMS error of 0.19ms−1 with a peak-to-peak noise of 0.09ms−1 and a range up to 28ms−1. Therefore, the probe fulfills the requirements for the planned purpose of active turbulence suppression for fixed-wing UAV. Furthermore, the probe appears also suitable for a variety of other applications.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"99 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116014390","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999721
Shaoming He, Hyo-Sang Shin, A. Tsourdos
This paper addresses the problem of energy-optimal waypoint-following guidance for an Unmanned Aerial Vehicle with the consideration of arbitrary number of obstacles. The proposed guidance law is derived as a solution of a linear quadratic optimal control problem in conjunction with convex parameter optimization. The algorithm developed integrates path following and obstacle avoidance into a single step and is able to be applied to a general waypoint-following mission. Several particular cases of the proposed guidance law are presented to provide better insights of the proposed algorithm. Nonlinear numerical simulations clearly demonstrate the effectiveness of the proposed formulations.
{"title":"Optimal Guidance for Integrated Waypoint Following and Obstacle Avoidance","authors":"Shaoming He, Hyo-Sang Shin, A. Tsourdos","doi":"10.1109/REDUAS47371.2019.8999721","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999721","url":null,"abstract":"This paper addresses the problem of energy-optimal waypoint-following guidance for an Unmanned Aerial Vehicle with the consideration of arbitrary number of obstacles. The proposed guidance law is derived as a solution of a linear quadratic optimal control problem in conjunction with convex parameter optimization. The algorithm developed integrates path following and obstacle avoidance into a single step and is able to be applied to a general waypoint-following mission. Several particular cases of the proposed guidance law are presented to provide better insights of the proposed algorithm. Nonlinear numerical simulations clearly demonstrate the effectiveness of the proposed formulations.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116314056","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999714
Belem Rojas, H. Abaunza, P. Castillo, I. Thouvenin, R. Lozano
In this paper, a teleoperation system for a quadrotor using a virtual environment is presented. A virtual aerial vehicle was introduced by applying the dynamic model of a quadrotor. An attitude controller was implemented for an operator to pilot the drone. The behavior of the virtual drone is communicated to a real vehicle through a mapping between the real and virtual spaces. This mapping is introduced in a robust quaternion-based controller ensures a safe flight in the real world. Simulations and experimental results are carried out to validate the proposal. With this idea of controlling a drone virtually, the performance of beginner and expert drone pilots is improved.
{"title":"Teleoperation of a drone based on a virtual environment*","authors":"Belem Rojas, H. Abaunza, P. Castillo, I. Thouvenin, R. Lozano","doi":"10.1109/REDUAS47371.2019.8999714","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999714","url":null,"abstract":"In this paper, a teleoperation system for a quadrotor using a virtual environment is presented. A virtual aerial vehicle was introduced by applying the dynamic model of a quadrotor. An attitude controller was implemented for an operator to pilot the drone. The behavior of the virtual drone is communicated to a real vehicle through a mapping between the real and virtual spaces. This mapping is introduced in a robust quaternion-based controller ensures a safe flight in the real world. Simulations and experimental results are carried out to validate the proposal. With this idea of controlling a drone virtually, the performance of beginner and expert drone pilots is improved.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126825970","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999691
Husam A. Foudeh, P. Luk, J. Whidborne
Research into autonomous control and behaviour of mobile vehicles has become more and more widespread. Unmanned aerial vehicles (UAVs) have seen an upsurge of interest and of the many UAVs available, the quadrotor has shown significant potential in monitoring and surveillance tasks. This paper examines the performance of iterative learning control (ILC) in gradient-based control that enhances a quadrotor’s controllability and stability during attitude control. It describes the development of the learning algorithms which exploit the repeated nature of the fault-finding task. Iterative learning control algorithms are derived and implemented on a quadrotor in a test bench. The proposed ILC algorithms on the quadrotor model are evaluated for system stability, convergence speed, and trajectory tracking error. Finally, the performance of the proposed algorithms is compared against a baseline performance of the PID control schemes.
{"title":"Quadrotor System Design for a 3 DOF platform based on Iterative Learning Control","authors":"Husam A. Foudeh, P. Luk, J. Whidborne","doi":"10.1109/REDUAS47371.2019.8999691","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999691","url":null,"abstract":"Research into autonomous control and behaviour of mobile vehicles has become more and more widespread. Unmanned aerial vehicles (UAVs) have seen an upsurge of interest and of the many UAVs available, the quadrotor has shown significant potential in monitoring and surveillance tasks. This paper examines the performance of iterative learning control (ILC) in gradient-based control that enhances a quadrotor’s controllability and stability during attitude control. It describes the development of the learning algorithms which exploit the repeated nature of the fault-finding task. Iterative learning control algorithms are derived and implemented on a quadrotor in a test bench. The proposed ILC algorithms on the quadrotor model are evaluated for system stability, convergence speed, and trajectory tracking error. Finally, the performance of the proposed algorithms is compared against a baseline performance of the PID control schemes.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126521234","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999688
Chaoying Niu, Liang Sun
A new nonlinear disturbance observer (NDO) based generalized model predictive control approach is developed for quadrotor attitude tracking missions subject to the multiple model uncertainties including parametric uncertainties, matched and mismatched external disturbances. Based on the kinematics and dynamics of the systems, the attitude tracking controller development is divided into two sub-systems designs to facilitate dealing with the matched and mismatched uncertainties. For the attitude kinematics subsystem, the generalized model predictive method and NDO are combined to design the desired angular velocity and compensate the mismatched disturbances. For the attitude dynamics subsystem, the generalized model predictive method and NDO are integrated again to design the actual control input torques of quadrotors and compensate the matched disturbances. In particular, NDO is used to estimate the unknown disturbances, while the model predictive control is used for optimizing the desired angular velocity and the control efforts. The attitude tracking errors eventually converge to small neighborhoods of zero with the proposed control strategy, and the simulation results validate the effectiveness of the proposed approach.
{"title":"Nonlinear Disturbance Observer Augmented Model Predictive Attitude Tracking Control of Quadrotors","authors":"Chaoying Niu, Liang Sun","doi":"10.1109/REDUAS47371.2019.8999688","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999688","url":null,"abstract":"A new nonlinear disturbance observer (NDO) based generalized model predictive control approach is developed for quadrotor attitude tracking missions subject to the multiple model uncertainties including parametric uncertainties, matched and mismatched external disturbances. Based on the kinematics and dynamics of the systems, the attitude tracking controller development is divided into two sub-systems designs to facilitate dealing with the matched and mismatched uncertainties. For the attitude kinematics subsystem, the generalized model predictive method and NDO are combined to design the desired angular velocity and compensate the mismatched disturbances. For the attitude dynamics subsystem, the generalized model predictive method and NDO are integrated again to design the actual control input torques of quadrotors and compensate the matched disturbances. In particular, NDO is used to estimate the unknown disturbances, while the model predictive control is used for optimizing the desired angular velocity and the control efforts. The attitude tracking errors eventually converge to small neighborhoods of zero with the proposed control strategy, and the simulation results validate the effectiveness of the proposed approach.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"198 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132322042","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999712
Shuoyuan Xu, Hyo-Sang Shin, A. Tsourdos
This paper proposes a novel clustering-based distributed multi-target tracking algorithm over a sensor network. Each local sensor runs a joint probabilistic data association filter to obtain local state estimation. The estimates are communicated between connected sensors for track-totrack association and fusion. A novel distributed DBSCAN (D-DBSCAN) clustering algorithm is proposed to solve the track-to-track association problem. The proposed algorithm shows advantages in computational efficiency compared with conventional distributed multi-target tracking approaches. Extensive simulations provided substantial evidence for the effectiveness of the proposed algorithm.
{"title":"Distributed Multi-Target Tracking with D-DBSCAN Clustering","authors":"Shuoyuan Xu, Hyo-Sang Shin, A. Tsourdos","doi":"10.1109/REDUAS47371.2019.8999712","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999712","url":null,"abstract":"This paper proposes a novel clustering-based distributed multi-target tracking algorithm over a sensor network. Each local sensor runs a joint probabilistic data association filter to obtain local state estimation. The estimates are communicated between connected sensors for track-totrack association and fusion. A novel distributed DBSCAN (D-DBSCAN) clustering algorithm is proposed to solve the track-to-track association problem. The proposed algorithm shows advantages in computational efficiency compared with conventional distributed multi-target tracking approaches. Extensive simulations provided substantial evidence for the effectiveness of the proposed algorithm.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"627 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113994334","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999697
Alejandro Castillejo-Calle, Jose A. Millan-Romera, Héctor Pérez-Leon, J. L. Andrade-Pineda, I. Maza, A. Ollero
This paper presents the first application of the multi-purpose ROS-MAGNA [1] framework to a real inspection mission performed with a team of multiple UAS. This framework makes transparent the type of autopilot on-board and creates the state machines that control the behaviour of the different UAS from the specification of the multi-UAS mission. In addition, it integrates a virtual world generation tool to manage the environment’s information and visualize the geometrical objects of interest to properly follow the progress of the mission. Due to its flexibility, it can be applied to the inspection of different photovoltaic plants with teams of UAS. The paper describes the system architecture and the human machine interface developed for the mission management by the personnel of the photovoltaic plant. Finally, we show both simulations results and real flight experiments on a photovoltaic plant located in Utrera (Spain).
{"title":"A multi-UAS system for the inspection of photovoltaic plants based on the ROS-MAGNA framework","authors":"Alejandro Castillejo-Calle, Jose A. Millan-Romera, Héctor Pérez-Leon, J. L. Andrade-Pineda, I. Maza, A. Ollero","doi":"10.1109/REDUAS47371.2019.8999697","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999697","url":null,"abstract":"This paper presents the first application of the multi-purpose ROS-MAGNA [1] framework to a real inspection mission performed with a team of multiple UAS. This framework makes transparent the type of autopilot on-board and creates the state machines that control the behaviour of the different UAS from the specification of the multi-UAS mission. In addition, it integrates a virtual world generation tool to manage the environment’s information and visualize the geometrical objects of interest to properly follow the progress of the mission. Due to its flexibility, it can be applied to the inspection of different photovoltaic plants with teams of UAS. The paper describes the system architecture and the human machine interface developed for the mission management by the personnel of the photovoltaic plant. Finally, we show both simulations results and real flight experiments on a photovoltaic plant located in Utrera (Spain).","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122859262","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999701
W. Ra, Hyo-Sang Shin, Yunha Lee, I. Whang
This paper proposes a novel time-to-go estimation filter necessary for implementing the precise vertical homing guidance law of an anti-ship missiles with a RF seeker. By investigating the behavior of a pure proportional navigation guidance, it is shown that the time-to-go can be expressed as a Gauss hypergeometric function of the relative range, closing velocity and heading error. As an effective way to solve the time-to-go estimation problem, a recursive time-to-go estimation scheme with two-stage filter structure is taken into account. The suggested approach is able to not only consider the length of the curved homing trajectory caused by the heading error but also mitigate the unwanted time-to-go estimation errors due to the correlations among seeker measurement noises. Through the simulations for typical anti-ship missile homing scenarios, it is demonstrated that our method could provide superior time-to-go estimation performance compared to the previous ones.
{"title":"Recursive Time-to-go Estimator for Anti-ship Missiles Guided by Pure Proportional Navigation","authors":"W. Ra, Hyo-Sang Shin, Yunha Lee, I. Whang","doi":"10.1109/REDUAS47371.2019.8999701","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999701","url":null,"abstract":"This paper proposes a novel time-to-go estimation filter necessary for implementing the precise vertical homing guidance law of an anti-ship missiles with a RF seeker. By investigating the behavior of a pure proportional navigation guidance, it is shown that the time-to-go can be expressed as a Gauss hypergeometric function of the relative range, closing velocity and heading error. As an effective way to solve the time-to-go estimation problem, a recursive time-to-go estimation scheme with two-stage filter structure is taken into account. The suggested approach is able to not only consider the length of the curved homing trajectory caused by the heading error but also mitigate the unwanted time-to-go estimation errors due to the correlations among seeker measurement noises. Through the simulations for typical anti-ship missile homing scenarios, it is demonstrated that our method could provide superior time-to-go estimation performance compared to the previous ones.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"184 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123870639","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 : 2019-11-01DOI: 10.1109/REDUAS47371.2019.8999679
On Park, Hyo-Sang Shin, A. Tsourdos
This paper describes a Reinforcement Learning (RL) application using Linear Quadratic Regulator (LQR) based tracking controller, which is augmented with a tracking error term. In order to deal with the steady-state errors, Linear Quadratic Tracker with Integrator (LQTI) is designed by adding an integration term of the tracking error in the state variable. Based on the LQTI, an online learning using the Integral Reinforcement Learning (IRL) is applied for the tracking problem to find the optimal control on the partially unknown continuous-time systems by regulating the augmented state variable. The optimal control solution and the performance of the method are verified through numerical simulation on two applications.
{"title":"Linear Quadratic Tracker with Integrator using Integral Reinforcement Learning","authors":"On Park, Hyo-Sang Shin, A. Tsourdos","doi":"10.1109/REDUAS47371.2019.8999679","DOIUrl":"https://doi.org/10.1109/REDUAS47371.2019.8999679","url":null,"abstract":"This paper describes a Reinforcement Learning (RL) application using Linear Quadratic Regulator (LQR) based tracking controller, which is augmented with a tracking error term. In order to deal with the steady-state errors, Linear Quadratic Tracker with Integrator (LQTI) is designed by adding an integration term of the tracking error in the state variable. Based on the LQTI, an online learning using the Integral Reinforcement Learning (IRL) is applied for the tracking problem to find the optimal control on the partially unknown continuous-time systems by regulating the augmented state variable. The optimal control solution and the performance of the method are verified through numerical simulation on two applications.","PeriodicalId":351115,"journal":{"name":"2019 Workshop on Research, Education and Development of Unmanned Aerial Systems (RED UAS)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128570729","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：