Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained Inputs

In this paper, the robust tracking control for unknown multi-player nonlinear systems with constrained inputs and uncertainties is studied by data-driven reinforcement learning scheme. In this study, the dependence on system dynamics is relaxed through the use of system input and state data. First o...

Full description

Saved in:
Bibliographic Details
Main Authors: Jingang Zhao, Jun Zhao, Yehan Chang, Guosheng Xu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11077137/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849318047806914560
author Jingang Zhao
Jun Zhao
Yehan Chang
Guosheng Xu
author_facet Jingang Zhao
Jun Zhao
Yehan Chang
Guosheng Xu
author_sort Jingang Zhao
collection DOAJ
description In this paper, the robust tracking control for unknown multi-player nonlinear systems with constrained inputs and uncertainties is studied by data-driven reinforcement learning scheme. In this study, the dependence on system dynamics is relaxed through the use of system input and state data. First of all, an augmented multi-player uncertain system is built by integrating the multi-player uncertain nonlinear systems and the given reference signal system. Afterwards, a non-quadratic cost function including the bounded functions, system states and all players control inputs is designed to handle the input constraints and uncertainties, and the robust tracking control problem is therefore converted into a constrained optimal control problem of the nominal system corresponding to the augmented multi-player uncertain system. Further, we rigorously analyze the equivalence of this conversion under specific conditions. The solution of the constrained optimal control problem can be obtained by solving a coupled Hamilton-Jacobi (HJ) equations. Then, on the basis of a given policy iteration scheme, a data-driven reinforcement learning scheme utilizing actor-critic neural networks is designed to solve the coupled HJ equation under the restriction of unavailable system dynamics. The weights of the neural network are learned using the least squares method from the input and state data collected from the system. Finally, a two-player uncertain nonlinear system and an induction heater circuit system are simulated to support the theoretical analysis.
format Article
id doaj-art-c47db1dfff2a4e909d834c3f001bf034
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-c47db1dfff2a4e909d834c3f001bf0342025-08-20T03:50:59ZengIEEEIEEE Access2169-35362025-01-011312106112107210.1109/ACCESS.2025.358780811077137Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained InputsJingang Zhao0https://orcid.org/0000-0002-2583-8446Jun Zhao1Yehan Chang2https://orcid.org/0009-0009-2860-4846Guosheng Xu3https://orcid.org/0009-0004-2602-9482School of Machinery and Automation, Weifang University, Weifang, Shandong, ChinaArtificial Intelligence Key Laboratory of Sichuan Province, School of Automation and Information Engineering, Sichuan University of Science and Engineering, Yibin, ChinaSchool of Machinery and Automation, Weifang University, Weifang, Shandong, ChinaGoertek College of Science and Technology Industry, Weifang University, Weifang, Shandong, ChinaIn this paper, the robust tracking control for unknown multi-player nonlinear systems with constrained inputs and uncertainties is studied by data-driven reinforcement learning scheme. In this study, the dependence on system dynamics is relaxed through the use of system input and state data. First of all, an augmented multi-player uncertain system is built by integrating the multi-player uncertain nonlinear systems and the given reference signal system. Afterwards, a non-quadratic cost function including the bounded functions, system states and all players control inputs is designed to handle the input constraints and uncertainties, and the robust tracking control problem is therefore converted into a constrained optimal control problem of the nominal system corresponding to the augmented multi-player uncertain system. Further, we rigorously analyze the equivalence of this conversion under specific conditions. The solution of the constrained optimal control problem can be obtained by solving a coupled Hamilton-Jacobi (HJ) equations. Then, on the basis of a given policy iteration scheme, a data-driven reinforcement learning scheme utilizing actor-critic neural networks is designed to solve the coupled HJ equation under the restriction of unavailable system dynamics. The weights of the neural network are learned using the least squares method from the input and state data collected from the system. Finally, a two-player uncertain nonlinear system and an induction heater circuit system are simulated to support the theoretical analysis.https://ieeexplore.ieee.org/document/11077137/Data-driven reinforcement learningrobust tracking controlmulti-player systemsneural networks (NNs)
spellingShingle Jingang Zhao
Jun Zhao
Yehan Chang
Guosheng Xu
Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained Inputs
IEEE Access
Data-driven reinforcement learning
robust tracking control
multi-player systems
neural networks (NNs)
title Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained Inputs
title_full Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained Inputs
title_fullStr Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained Inputs
title_full_unstemmed Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained Inputs
title_short Data-Driven Robust Tracking Control for Multi-Player Nonzero-Sum Games With Constrained Inputs
title_sort data driven robust tracking control for multi player nonzero sum games with constrained inputs
topic Data-driven reinforcement learning
robust tracking control
multi-player systems
neural networks (NNs)
url https://ieeexplore.ieee.org/document/11077137/
work_keys_str_mv AT jingangzhao datadrivenrobusttrackingcontrolformultiplayernonzerosumgameswithconstrainedinputs
AT junzhao datadrivenrobusttrackingcontrolformultiplayernonzerosumgameswithconstrainedinputs
AT yehanchang datadrivenrobusttrackingcontrolformultiplayernonzerosumgameswithconstrainedinputs
AT guoshengxu datadrivenrobusttrackingcontrolformultiplayernonzerosumgameswithconstrainedinputs