Preston Fairchild, Noah Shepard, Yu Mei, Xiaobo Tan
The inherent low stiffness in soft robots makes them preferable for working in close proximity to humans. However, having this low stiffness creates challenges when operating in terms of control and sensitivity to disturbances. To alleviate this issue, soft robots often have built-in stiffness tuning mechanisms that allow for controlled increases in stiffness. Additionally, redundant pneumatic manipulators can utilize antagonistic pressure to achieve identical positions under increased stiffness. In this paper, we develop a model to predict the stiffness and configuration of a pneumatic soft manipulator under different pressure inputs and external forces. The model is developed based on the physical characteristics of a soft manipulator while enabling efficient parameter estimation and computation. The efficacy of the modeling approach is supported via experimental results.
{"title":"Semi-physical Modeling of Soft Pneumatic Actuators with Stiffness Tuning","authors":"Preston Fairchild, Noah Shepard, Yu Mei, Xiaobo Tan","doi":"10.1115/1.4064090","DOIUrl":"https://doi.org/10.1115/1.4064090","url":null,"abstract":"The inherent low stiffness in soft robots makes them preferable for working in close proximity to humans. However, having this low stiffness creates challenges when operating in terms of control and sensitivity to disturbances. To alleviate this issue, soft robots often have built-in stiffness tuning mechanisms that allow for controlled increases in stiffness. Additionally, redundant pneumatic manipulators can utilize antagonistic pressure to achieve identical positions under increased stiffness. In this paper, we develop a model to predict the stiffness and configuration of a pneumatic soft manipulator under different pressure inputs and external forces. The model is developed based on the physical characteristics of a soft manipulator while enabling efficient parameter estimation and computation. The efficacy of the modeling approach is supported via experimental results.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"46 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139264452","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}
Pouria Karimi Shahri, B. Homchaudhuri, Azad Ghaffari, Amir Ghasemi
The primary aim of this research paper is to enhance the effectiveness of a two-level infrastructure-based control framework utilized for traffic management in expansive networks. The lower-level controller adjusts vehicle velocities to achieve the desired density determined by the upper-level controller. The upper-level controller employs a novel Lyapunov- based switched Newton extremum-seeking control approach to ascertain the optimal vehicle density in congested cells where downstream bottlenecks are unknown, even in the presence of disturbances in the model. Unlike gradient-based approaches, the Newton algorithm eliminates the need for the unknown Hessian matrix, allowing for user-assignable convergence rates. The Lyapunov-based switched approach also ensures asymptotic convergence to the optimal set point. Simulation results demonstrate that the proposed approach, combining Newton's method with user-assignable convergence rates and a Lyapunov-based switch, outperforms gradient based extremum-seeking in the hierarchical control framework.
{"title":"Improving the Performance of a Hierarchical Traffic Flow Control Framework using Lyapunov-based Switched Newton Extremum Seeking","authors":"Pouria Karimi Shahri, B. Homchaudhuri, Azad Ghaffari, Amir Ghasemi","doi":"10.1115/1.4064088","DOIUrl":"https://doi.org/10.1115/1.4064088","url":null,"abstract":"The primary aim of this research paper is to enhance the effectiveness of a two-level infrastructure-based control framework utilized for traffic management in expansive networks. The lower-level controller adjusts vehicle velocities to achieve the desired density determined by the upper-level controller. The upper-level controller employs a novel Lyapunov- based switched Newton extremum-seeking control approach to ascertain the optimal vehicle density in congested cells where downstream bottlenecks are unknown, even in the presence of disturbances in the model. Unlike gradient-based approaches, the Newton algorithm eliminates the need for the unknown Hessian matrix, allowing for user-assignable convergence rates. The Lyapunov-based switched approach also ensures asymptotic convergence to the optimal set point. Simulation results demonstrate that the proposed approach, combining Newton's method with user-assignable convergence rates and a Lyapunov-based switch, outperforms gradient based extremum-seeking in the hierarchical control framework.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"5 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139262661","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}
Abstract We propose a new theory for fluid-structure interactions of cantilever microbeams undergoing small amplitude vibrations in viscous fluids. The method is based on the concept of nonlocal modal hydrodynamic functions that accurately capture 3D fluid loading on the structure. For short beams for which 3D effects become prominent, existing local theories based on 2D fluid approximations are inadequate to predict the dynamic response. We discuss and compare model predictions in terms of frequency response functions, modal shapes, quality factors, and added mass ratios with the predictions of the local theory, and validate our new model with experimental results.
{"title":"Nonlocal theory for submerged cantilever beams undergoing torsional vibrations","authors":"Burak Gulsacan, Matteo Aureli","doi":"10.1115/1.4063994","DOIUrl":"https://doi.org/10.1115/1.4063994","url":null,"abstract":"Abstract We propose a new theory for fluid-structure interactions of cantilever microbeams undergoing small amplitude vibrations in viscous fluids. The method is based on the concept of nonlocal modal hydrodynamic functions that accurately capture 3D fluid loading on the structure. For short beams for which 3D effects become prominent, existing local theories based on 2D fluid approximations are inadequate to predict the dynamic response. We discuss and compare model predictions in terms of frequency response functions, modal shapes, quality factors, and added mass ratios with the predictions of the local theory, and validate our new model with experimental results.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"48 6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135584751","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}
Hyunjin Ahn, Heran Shen, Xingyu Zhou, Yung-Chi Kung, Junmin Wang
Abstract Accurate estimation of the state of charge (SOC) is crucial for ensuring the safe and efficient operation of lithium-ion batteries. Machine learning (ML) models may achieve high SOC estimation accuracy, but typically require large training datasets that may not always be accessible in practical applications. To address this issue, this work proposes a hybrid model consisting of a Transformer neural network and a single particle model with electrolyte dynamics (SPMe) for SOC estimation in limited data scenarios. The Transformer can leverage the internal battery states estimated by the SPMe when necessary and learn to use information from multiple sources (i.e., experimental data and SPMe). Two limited data scenarios, partially available cycles and varying temperatures, are evaluated with experimental battery discharge cycles to identify the conditions under which the proposed model outperforms traditional ML models. Despite being highly dependent on the SPMe's performance, the hybrid model demonstrated improved SOC estimation over the baseline models, with less than 2% error for most scenarios.
{"title":"State of Charge Estimation of Lithium-ion Batteries using Physics-informed Transformer for Limited Data Scenarios","authors":"Hyunjin Ahn, Heran Shen, Xingyu Zhou, Yung-Chi Kung, Junmin Wang","doi":"10.1115/1.4063995","DOIUrl":"https://doi.org/10.1115/1.4063995","url":null,"abstract":"Abstract Accurate estimation of the state of charge (SOC) is crucial for ensuring the safe and efficient operation of lithium-ion batteries. Machine learning (ML) models may achieve high SOC estimation accuracy, but typically require large training datasets that may not always be accessible in practical applications. To address this issue, this work proposes a hybrid model consisting of a Transformer neural network and a single particle model with electrolyte dynamics (SPMe) for SOC estimation in limited data scenarios. The Transformer can leverage the internal battery states estimated by the SPMe when necessary and learn to use information from multiple sources (i.e., experimental data and SPMe). Two limited data scenarios, partially available cycles and varying temperatures, are evaluated with experimental battery discharge cycles to identify the conditions under which the proposed model outperforms traditional ML models. Despite being highly dependent on the SPMe's performance, the hybrid model demonstrated improved SOC estimation over the baseline models, with less than 2% error for most scenarios.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135636226","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}
Abstract Linear stability of a car following model is investigated considering time delays due to human reaction times and communication delays. In multiple trials, we generate networks of randomly arranged legacy and autonomous vehicles, each vehicle type respecting some connectivity rules. Next, the largest amount of delay the vehicle network can tolerate without becoming unstable is computed. This delay, also known as the delay margin (DM), can become sensitive with respect to platoon size under certain connectivity rules. We demonstrate this ‘fragility’ property, and next propose a new network-design policy with which DM can be made robust against platoon size.
{"title":"Fragility of Delay Margin Against Platoon Size in a Connected Vehicle System; Network Design and Multiple Trials","authors":"Duo Wang, Rifat Sipahi","doi":"10.1115/1.4063657","DOIUrl":"https://doi.org/10.1115/1.4063657","url":null,"abstract":"Abstract Linear stability of a car following model is investigated considering time delays due to human reaction times and communication delays. In multiple trials, we generate networks of randomly arranged legacy and autonomous vehicles, each vehicle type respecting some connectivity rules. Next, the largest amount of delay the vehicle network can tolerate without becoming unstable is computed. This delay, also known as the delay margin (DM), can become sensitive with respect to platoon size under certain connectivity rules. We demonstrate this ‘fragility’ property, and next propose a new network-design policy with which DM can be made robust against platoon size.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134977176","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}
Bo Ying Su, Zhongqi Wei, James McCann, Wenzhen Yuan, Changliu Liu
Abstract Textiles are a promising material for building tactile skins because they are lightweight, flexible, and can conform to complex surfaces. They enhance robot-human interaction by localizing contact points and measuring contact forces. This paper presents a solution for rapidly fabricating, calibrating, and deploying these skins on industrial robot arms. The novel automated skin calibration procedure maps skin locations to robot geometry and calibrates contact force. Through experiments on a FANUC LR Mate 200id/7L industrial robot, we demonstrate that tactile skins made from textiles can be effectively used for human-robot interaction in industrial environments. We demonstrated the following: 1) modifying trajectories using human contact signals and 2) improving force interactions with humans using tactile skins This work presents a comprehensive solution for rapidly fabricating, calibrating, and deploying textile and tactile skins for interactive industrial robots. Our contributions are as follows: 1) we introduce a method for determining skin parameters based on robot geometry; 2) we propose an automated skin calibration process that calibrates both the position and force reading from the skin to the robot; and 3) we demonstrate how the skins can be integrated into robot control and learning to improve the safety and interactivity of industrial robots through experiments on a FANUC LR Mate 200id/7L industrial robot arm.
纺织品是一种很有前途的材料，因为它们重量轻，柔韧性好，可以适应复杂的表面。它们通过定位接触点和测量接触力来增强人机交互。本文提出了一种在工业机器人手臂上快速制造、校准和部署这些皮肤的解决方案。新的自动皮肤校准程序映射皮肤位置到机器人的几何形状和校准接触力。通过在FANUC LR Mate 200id/7L工业机器人上的实验，我们证明了由纺织品制成的触觉皮肤可以有效地用于工业环境中的人机交互。我们演示了以下内容:1)使用人体接触信号修改轨迹;2)使用触觉皮肤改善与人类的力交互。这项工作为交互式工业机器人快速制造、校准和部署纺织品和触觉皮肤提供了一个全面的解决方案。我们的贡献如下:1)我们介绍了一种基于机器人几何形状确定皮肤参数的方法;2)我们提出了一种自动皮肤校准过程，可以校准从皮肤到机器人的位置和力读数;3)通过在FANUC LR Mate 200id/7L工业机器人手臂上的实验，我们展示了如何将皮肤集成到机器人控制和学习中，以提高工业机器人的安全性和交互性。
{"title":"CUSTOMIZING TEXTILE AND TACTILE SKINS FOR INTERACTIVE INDUSTRIAL ROBOTS","authors":"Bo Ying Su, Zhongqi Wei, James McCann, Wenzhen Yuan, Changliu Liu","doi":"10.1115/1.4063602","DOIUrl":"https://doi.org/10.1115/1.4063602","url":null,"abstract":"Abstract Textiles are a promising material for building tactile skins because they are lightweight, flexible, and can conform to complex surfaces. They enhance robot-human interaction by localizing contact points and measuring contact forces. This paper presents a solution for rapidly fabricating, calibrating, and deploying these skins on industrial robot arms. The novel automated skin calibration procedure maps skin locations to robot geometry and calibrates contact force. Through experiments on a FANUC LR Mate 200id/7L industrial robot, we demonstrate that tactile skins made from textiles can be effectively used for human-robot interaction in industrial environments. We demonstrated the following: 1) modifying trajectories using human contact signals and 2) improving force interactions with humans using tactile skins This work presents a comprehensive solution for rapidly fabricating, calibrating, and deploying textile and tactile skins for interactive industrial robots. Our contributions are as follows: 1) we introduce a method for determining skin parameters based on robot geometry; 2) we propose an automated skin calibration process that calibrates both the position and force reading from the skin to the robot; and 3) we demonstrate how the skins can be integrated into robot control and learning to improve the safety and interactivity of industrial robots through experiments on a FANUC LR Mate 200id/7L industrial robot arm.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135696198","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}
Abstract Stabilizing a wafer's temperature during plasma etching is a critical issue in semiconductor manufacturing. In this study, we propose feedback control of the wafer temperature using the pressure of helium gas (He) that is fed into the gap between the wafer and an electrostatic chuck (ESC) and an algorithm of the model predictive control (MPC) combined with an observer. The temperatures are measured only at the wafer edge zone and the ESC ceramic plate that are accessible during the process. The observer estimates wafer temperatures of center and edge zones and the injected heat power with the help of the measured edge zone temperature. The MPC determines the optimal He pressures based on the estimated temperatures to control both zone temperatures. The algorithm of the feedback control was formulated, and its validity was experimentally confirmed. Results showed that the observer worked well to estimate both zone wafer temperatures and the injected heat power. Results also showed that the temperatures were successfully controlled.
{"title":"WAFER TEMPERATURE CONTROL USING HELIUM PRESSURE AND OBSERVER-BASED MPC","authors":"Daisuke Hayashi, Kotaro Takijiri, Takayuki Ueda","doi":"10.1115/1.4063600","DOIUrl":"https://doi.org/10.1115/1.4063600","url":null,"abstract":"Abstract Stabilizing a wafer's temperature during plasma etching is a critical issue in semiconductor manufacturing. In this study, we propose feedback control of the wafer temperature using the pressure of helium gas (He) that is fed into the gap between the wafer and an electrostatic chuck (ESC) and an algorithm of the model predictive control (MPC) combined with an observer. The temperatures are measured only at the wafer edge zone and the ESC ceramic plate that are accessible during the process. The observer estimates wafer temperatures of center and edge zones and the injected heat power with the help of the measured edge zone temperature. The MPC determines the optimal He pressures based on the estimated temperatures to control both zone temperatures. The algorithm of the feedback control was formulated, and its validity was experimentally confirmed. Results showed that the observer worked well to estimate both zone wafer temperatures and the injected heat power. Results also showed that the temperatures were successfully controlled.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135696203","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}
Abstract This paper studies the combined maneuver of flying and sailing for a robotic system which is referred to as a flying+sailing drone. Due to the emergence of hybrid systems behavior in tasks which involve both the flying and sailing modes, a hybrid systems formulation of the robotic system is presented. Key characteristics of the system are (i) changes in the dimension of the state space as the system switches from flying to sailing and vice versa and (ii) the presence of autonomous switchings triggered only upon the landing of the drone on the water surface. For the scenario in which the drone's initial state is given in the flying mode and a fixed terminal state is specified in the sailing mode, the associated optimal control problems are studied within the vertical plane passing through the given points, hence the dynamics of the drone in the flying mode are represented in a five-dimensional state space (associated with three degrees-of-freedom) and in a three-dimensional state space in the sailing mode (associated with two degrees-of-freedom). In particular, the optimal control problems for the minimization of time and the minimization of the control effort are formulated, the associated necessary optimality conditions are obtained from the Hybrid Minimum Principle (HMP), and the associated numerical simulations are presented.
{"title":"Hybrid Optimal Control of a Flying+Sailing Drone","authors":"Taha Yasini, Ali Pakniyat","doi":"10.1115/1.4063603","DOIUrl":"https://doi.org/10.1115/1.4063603","url":null,"abstract":"Abstract This paper studies the combined maneuver of flying and sailing for a robotic system which is referred to as a flying+sailing drone. Due to the emergence of hybrid systems behavior in tasks which involve both the flying and sailing modes, a hybrid systems formulation of the robotic system is presented. Key characteristics of the system are (i) changes in the dimension of the state space as the system switches from flying to sailing and vice versa and (ii) the presence of autonomous switchings triggered only upon the landing of the drone on the water surface. For the scenario in which the drone's initial state is given in the flying mode and a fixed terminal state is specified in the sailing mode, the associated optimal control problems are studied within the vertical plane passing through the given points, hence the dynamics of the drone in the flying mode are represented in a five-dimensional state space (associated with three degrees-of-freedom) and in a three-dimensional state space in the sailing mode (associated with two degrees-of-freedom). In particular, the optimal control problems for the minimization of time and the minimization of the control effort are formulated, the associated necessary optimality conditions are obtained from the Hybrid Minimum Principle (HMP), and the associated numerical simulations are presented.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135696140","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}
Lance McCann, Yoshua Gombo, Anuj Tiwari, Joseph Garbini, Santosh Devasia
Abstract In manufacturing operations such as clamping and drilling of elastic structures, tool-workpiece normality must be maintained, and shear forces minimized to avoid tool or workpiece damage. The challenge is that the combined stiffness of a robot and workpiece, needed to control the robot-workpiece elastic interactions are often difficult to model and can vary due to geometry changes of the workpiece caused by large deformations and associated pose variations of the robot. The main contribution of this article is an algorithm, (i) to learn the robot-workpiece stiffness relationship using a model-free databased approach and (ii) to use it for applying desired forces and torques on the elastic structure. Moreover, comparative experiments with and without the data-based stiffness estimation show that clamping operating speed is increased by four times when using the stiffness estimation method while interaction forces and torques are kept within acceptable bounds.
{"title":"Data-based Stiffness Estimation for Control of Robot-Workpiece Elastic Interactions","authors":"Lance McCann, Yoshua Gombo, Anuj Tiwari, Joseph Garbini, Santosh Devasia","doi":"10.1115/1.4063606","DOIUrl":"https://doi.org/10.1115/1.4063606","url":null,"abstract":"Abstract In manufacturing operations such as clamping and drilling of elastic structures, tool-workpiece normality must be maintained, and shear forces minimized to avoid tool or workpiece damage. The challenge is that the combined stiffness of a robot and workpiece, needed to control the robot-workpiece elastic interactions are often difficult to model and can vary due to geometry changes of the workpiece caused by large deformations and associated pose variations of the robot. The main contribution of this article is an algorithm, (i) to learn the robot-workpiece stiffness relationship using a model-free databased approach and (ii) to use it for applying desired forces and torques on the elastic structure. Moreover, comparative experiments with and without the data-based stiffness estimation show that clamping operating speed is increased by four times when using the stiffness estimation method while interaction forces and torques are kept within acceptable bounds.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135696197","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}
Da Zhang, Junaid Anwar, Syed Ali Asad Rizvi, Yusheng Wei
Abstract We propose a novel deep learning-based approach for the problem of continuous-time leader synchronization in graphical games on large networks. The problem setup is to deploy a distributed and coordinated swarm to track the trajectory of a leader while minimizing local neighborhood tracking error and control cost for each agent. The goal of our work is to develop optimal control poli- cies for continuous-time leader synchronization in graphical games using deep neural networks. We discretize the agents model using sampling to facilitate the modification of gradient descent methods for learning optimal control policies. The distributed swarm is deployed for a certain amount of time while keeping the control input of each agent constant during each sampling period. After collecting state and input data at each sampling time during one iteration, we update the weights of a deep neural network for each agent using collected data to minimize a loss function that characterizes the agents local neighborhood tracking error and the control cost. A modified gradient descent method is presented to overcome existing limitations. The performance of the proposed method is compared with two reinforcement learning-based methods in terms of robustness to initial neural network weights and initial local neighbor- hood tracking errors, and the scalability to networks with a large number of agents. Our approach has been shown to achieve superior performance compared with the other two methods.
{"title":"Deep Learning for Continuous-time Leader Synchronization in Graphical Games Using Sampling and Deep Neural Networks","authors":"Da Zhang, Junaid Anwar, Syed Ali Asad Rizvi, Yusheng Wei","doi":"10.1115/1.4063607","DOIUrl":"https://doi.org/10.1115/1.4063607","url":null,"abstract":"Abstract We propose a novel deep learning-based approach for the problem of continuous-time leader synchronization in graphical games on large networks. The problem setup is to deploy a distributed and coordinated swarm to track the trajectory of a leader while minimizing local neighborhood tracking error and control cost for each agent. The goal of our work is to develop optimal control poli- cies for continuous-time leader synchronization in graphical games using deep neural networks. We discretize the agents model using sampling to facilitate the modification of gradient descent methods for learning optimal control policies. The distributed swarm is deployed for a certain amount of time while keeping the control input of each agent constant during each sampling period. After collecting state and input data at each sampling time during one iteration, we update the weights of a deep neural network for each agent using collected data to minimize a loss function that characterizes the agents local neighborhood tracking error and the control cost. A modified gradient descent method is presented to overcome existing limitations. The performance of the proposed method is compared with two reinforcement learning-based methods in terms of robustness to initial neural network weights and initial local neighbor- hood tracking errors, and the scalability to networks with a large number of agents. Our approach has been shown to achieve superior performance compared with the other two methods.","PeriodicalId":327130,"journal":{"name":"ASME Letters in Dynamic Systems and Control","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135696208","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：