Neural networks optimized learning control of state constraints systems
Introduction
There are several methods to handle the nonlinear systems where adaptive neural network and fuzzy control are two hot research topics [1], [2], [3], [4], [5], [6], [7], [8]. During the past decades, the adaptive backstepping learning control design of strict feedback nonlinear systems has brought about much attention, and some representative research results have been reported, for example, [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16]. Refs. [6], [7], [8], [9] investigated the adaptive NN/fuzzy backstepping control problem for single-input and single-output (SISO) strict-feedback nonlinear systems. The authors in [10], [11], [12], [13], [14], [15], [16] developed adaptive NN/fuzzy backstepping control schemes for multi-input and multi-output (MIMO) nonlinear systems or interconnected large-scale nonlinear systems. Although a great progress has been made in adaptive NN/fuzzy backstepping control for strict-feedback nonlinear systems, the control problems on output or state constraint are not been considered in the above mentioned results. The output or state constraints are ubiquitous in physical systems, e.g., robot systems, aerospace system, and industrial control systems. Violation of these constraints during operation may result in performance degradation, hazards or system damage. Therefore, the control design for strict-feedback nonlinear systems with output or state constraints is very important. Recently, some adaptive NN/fuzzy control approaches have been proposed for strict-feedback nonlinear systems with output or state constraints [17], [18], [19], [20], [21], [22], [23]. The authors in [17], [18], [19], [20] investigated the adaptive NN/fuzzy output constraint control problem for strict feedback nonlinear systems, and the authors in [21], [22], [23] proposed adaptive NN/fuzzy control methods for strict feedback nonlinear systems with state constraints. It should be noted that the above mentioned adaptive neural network and fuzzy backstepping constraint control methods do not address the optimal control problem.
It is well known that optimal control problem is one of key issues in modern control theory. As for nonlinear systems, the optimal control problem eventually boils down to solve the equation of Hamilton–Jacobi-Bellman (HJB). However, since the HJB equation is a nonlinear partial differential equation, it is very difficult to obtain its closed-loop form solution. More recently, some adaptive neural networks or fuzzy optimal control schemes have been proposed nonlinear systems [24], [25], [26], [27], [28], [29], [30], [31], [32]. However, the controlled systems in [24], [25], [26], [27], [28], [29], [30], [31], [32] are not strict feedback nonlinear systems. Besides, the control design methods in [24], [25], [26], [27], [28], [29], [30], [31], [32] are not based on the backstepping design principle. The authors in [33], [34], [35], [36], [37], [38] proposed the adaptive NN/fuzzy backstepping optimal control methods for different types of strict feedback nonlinear systems. However, the above adaptive NN/fuzzy optimal control approaches all do not study the state constraint optimal control problem. To the best of our knowledge, there are no results on adaptive NN/fuzzy optimal control for strict feedback nonlinear systems will full states constraints.
Motivated by the above observations, this paper investigates the adaptive NN backstepping optimal learning tracking control problem for a class of strict-feedback nonlinear systems with state constraints and unknown nonlinear dynamics. NNs are used to approximate the unknown nonlinear functions, critic and actor networks, and an adaptive NN state identifier is constructed. By constructing Barrier Lyapunov functions and optimal Barrier type cost functions, and employing the actor-critic architecture, the virtual and actual optimal controllers are developed by the backstepping recursive control design algorithm. The proposed adaptive NN optimal control scheme can guarantee that the closed-loop system is uniformly ultimately bounded and the system output can tracking the given reference signal. Moreover, the states of the controlled system are ensured not to transgress their constrained sets. The main contributions of this paper are summarized in the following two aspects:
- 1.
This paper first investigates the NN backstepping optimal learning tracking control problem for strict-feedback nonlinear systems with unknown nonlinear dynamics. The proposed control scheme not only can guarantee the closed-loop system to be UUB, but also achieve the tracking control objective. Although [38] investigated the NN backstepping optimal control problem for strict-feedback nonlinear systems, the nonlinear dynamics in [38] are known.
- 2.
This paper first studies the adaptive NN backstepping optimal constraint control problem for strict-feedback nonlinear systems, and a NN state feedback optimal control method has been proposed, and the proposed method can guarantee that the states of the controlled system not to transgress their constrained sets. Although [33], [34], [35], [36], [37], [38] also investigated the intelligent optimal control problem, they do not consider the state constrained problem.
The rest of this work is organized as follows. Section 2 exhibits preliminaries for SISO strict-feedback nonlinear system. The neural network state identifier design and based-identifier optimized controller and stability analysis is exhibited in Section 4. Section 5 gives simulation results. Section 6 concludes this paper.
Section snippets
Nonlinear control problem
Consider the following SISO strict-feedback nonlinear systemwhere is the system state vector, is the control input, is the output of system, with are unknown smooth functions. and are assumed Lipschitz continuous and stabilizable on the sets containing origin. All the states are supposed to be constrained in the compact sets, i.e., , where is a known
NN controller design and stability analysis
In this section, a NN adaptive backstepping optimized controller is proposed based on the actor-critic architecture, and the stability analysis of the closed-loop system is given.
Simulation example
In this section, a numerical example is given to illustrate the effectiveness of the proposed control scheme.
Consider the following strict-feedback nonlinear system:where and are the system states and is control input, and , the reference signal is .
There exist six RBF neural networks in the control system. Neural networks contains five nodes with centers evenly spaced
Conclusion
In this study, a NN learning optimized tracking control approach has been proposed for a class of strict-feedback nonlinear systems with state constraints. The neural networks have been used to learn the unknown nonlinear dynamics, and a NN state identifier has been proposed. By constructing Barrier Lyapunov functions and optimal Barrier type cost functions for identifier error dynamic systems and the subsystems of the identifier systems, and adopting the backstepping recursive control design
CRediT authorship contribution statement
Xiaomei Li: Writing - original draft, Validation. Yongming Li: Methodology, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work is supported in part by the National Nature Science Foundation of China (No: 52070091), the Social Science Planning Foundation of Liaoning Province (No: L19BJY028), the Innovation Talent Support Plan Project of Liaoning Province in 2019.
Xiaomei Li received the B.S degree in technology economic management and the M.S degree in Business management from Liaoning University of Technology, Jinzhou, china, and received the PH.D degree in technology economic from Liaoning University, Shenyang, China. She is a professor and a Doctoral supervisor who had been as a scholar to visit Edith Cowan University in Australia 2015. Her current research interests include technology innovation and input-output efficiency evaluation for industry
References (38)
- et al.
Adaptive dynamic surface control of nonlinear systems with unknown dead zone in pure feedback form
Automatica
(2008) - et al.
Adaptive neural dynamic surface control of strict-feedback nonlinear systems with full state constraints and unmodeled dynamics
Automatica
(2017) - et al.
Neural network approach to continuous time direct adaptive optimal control for partially unknown nonlinear systems
Neural Netw.
(2009) - et al.
Data-based adaptive neural network optimal output feedback control for nonlinear systems with actuator saturation
Neurocomputing
(2017) - L.K. Wang, H.K. Lam, Further study on observer design for continuous-time Takagi-Sugeno fuzzy model with unknown...
- et al.
Finite-time adaptive fault-tolerant control for nonlinear systems with multiple faults
IEEE/CAA J. Autom. Sinica
(2019) - L.K. Wang, J.J. Liu, H.K. Lam, Further study on stabilization for continuous-time Takagi- Sugeno fuzzy systems with...
- et al.
Distributed containment maneuvering of multiple marine vessels via neurodynamics-based output feedback
IEEE T. Ind. Electron.
(2017) - et al.
Approximation-based adaptive fuzzy control for a class of non-strict-feedback stochastic nonlinear systems
Sci. China Inform. Sci.
(2014) - et al.
Observer-based adaptive fuzzy tracking control for strict-feedback nonlinear systems with unknown control gain functions
IEEE Trans. Cybern.
(2020)
Adaptive neural control for a class of pure-feedback nonlinear systems via dynamic surface technique
IEEE Trans. Neural Netw. Learn. Syst.
Globally stable adaptive backstepping neural network control for uncertain strict-feedback systems with tracking accuracy known a priori
IEEE Trans. Neural Netw. Learn. Syst.
Observer-based adaptive fuzzy backstepping dynamic surface control for a class of MIMO nonlinear systems
IEEE Trans. Syst., Man Cybern. B, Cybern.
Robust adaptive neural network control for a class of uncertain MIMO nonlinear systems with input nonlinearities
IEEE Trans. Neural Netw.
Observer-based adaptive fuzzy backstepping output feedback control of uncertain MIMO pure-feedback nonlinear systems
IEEE Trans. Fuzzy Syst.
Neural-network-based decentralized adaptive control of interconnected a class of large-scale nonlinear systems with time-delays
IEEE Trans. Syst., Man Cybern. B, Cybern.
Observed-based adaptive fuzzy decentralized tracking control for switched uncertain nonlinear large-scale systems with dead zones
IEEE Trans. Syst. Man Cybern. Syst.
Globally decentralized adaptive backstepping neural network tracking output control for unknown nonlinear interconnectioned systems
Asian J. Control
Adaptive neural network decentralized backstepping output-feedback control for nonlinear large-scale systems with time delays
IEEE Trans. Neural Netw.
Cited by (0)
Xiaomei Li received the B.S degree in technology economic management and the M.S degree in Business management from Liaoning University of Technology, Jinzhou, china, and received the PH.D degree in technology economic from Liaoning University, Shenyang, China. She is a professor and a Doctoral supervisor who had been as a scholar to visit Edith Cowan University in Australia 2015. Her current research interests include technology innovation and input-output efficiency evaluation for industry and enterprise.
Yongming Li received the B.S. degree and the M.S. degree in Applied Mathematics from Liaoning University of Technology, Jinzhou, China, in 2004 and 2007, respectively. He received the Ph.D degree in Transportation Information Engineering & Control from Dalian Maritime University, Dalian, China in 2014. He is currently a Professor in the College of Science, Liaoning University of Technology. His current research interests include adaptive control, fuzzy control and neural networks control for nonlinear systems.