Adaptive Neural Tracking Control for Discrete-Time Switched Nonlinear Systems with Dead Zone Inputs

This paper considers the problem of adaptive neural control of stochastic nonlinear systems in nonstrict-feedback form with unknown backlash-like hysteresis nonlinearities. To overcome the design difficulty of nonstrict-feedback structure, variable separation technique is used to decompose the unknown functions of all state variables into a sum of smooth functions of each error dynamic. By combining radial basis function neural networks' universal approximation capability with an adaptive backstepping technique, an adaptive neural control algorithm is proposed. It is shown that the proposed controller guarantees that all the signals in the closed-loop system are four-moment semiglobally uniformly ultimately bounded, and the tracking error eventually converges to a small neighborhood of the origin in the sense of mean quartic value. Simulation results further show the effectiveness of the presented control scheme.

In this paper, an adaptive control strategy is developed for discrete‐time switched nonlinear systems with actuator faults and dead‐zone input under arbitrary switching conditions. The actuator faults considered include loss‐of‐effectiveness and bias faults, which are unknown but bounded. The complex structure of these systems, combined with actuator faults and dead‐zone inputs, presents significant challenges for control and this problem is addressed by approximating the unknown functions of each subsystem using radial basis function neural networks. Under arbitrary switching signals, the suggested controller and adaptive laws ensure that all signals remain bounded and the system output tracks the reference signal with a small bounded tracking error. The efficiency of the control technique is proven by two numerical examples.

In this paper, an novel adaptive tracking control is developed for a class of completely non-affine pure-feedback nonlinear systems using radial basis function neural networks (RBFNNs). Combining the dynamic surface control (DSC) technique and backstepping method, the explosion of complexity in the traditional backstepping design is avoided. Using mean value theorem and Young's inequality, only one learning parameter need to be tuned online in the whole controller design, and the computational burden is effectively alleviated. It is proved that the proposed design method is able to guarantee semi-global uniform ultimate boundedness (SGUUB) of all signals in the closed-loop system. Simulation results verify the effectiveness of the proposed approach.

This paper focuses on the problem of the adaptive neural control for a class of a perturbed pure-feedback nonlinear system. Based on radial basis function (RBF) neural networks’ universal approximation capability, an adaptive neural controller is developed via the backstepping technique. The proposed controller guarantees that all the signals in the closed-loop system are bounded and the tracking error eventually converges to a small neighborhood around the origin. The main advantage of this note lies in that a control strategy is presented for a class of pure-feedback nonlinear systems with external disturbances being bounded by functions of all state variables. A numerical example is provided to illustrate the effectiveness of the suggested approach.

This paper focuses on the problem of adaptive output-constrained neural tracking control for uncertain nonstrict-feedback systems in the presence of unknown symmetric output dead zone and input saturation. A Nussbaum-type function-based dead-zone model is introduced such that the dynamic surface control approach can be used for controller design. The variable separation technique is employed to decompose the unknown function of entire states in each subsystem into a series of smooth functions. Radial basis function neural networks are utilized to approximate the unknown black-box functions derived from Young's inequality. With the help of auxiliary first-order filters, the dimensions of neural network input are reduced in each recursive design. A main advantage of the proposed method is that for an -order nonlinear system, only one adaptation parameter needs to be tuned online. It is rigorously shown that the proposed output-constrained controller guarantees that all the closed-loop signals are semiglobal uniformly ultimately bounded and the tracking error never violates the output constraint.

Adaptive neural network control for nonlinear output-feedback systems under disturbances with unknown bounds

This study presents the disturbance observer-based adaptive neural tracking control approach for non-linear systems in non-strict-feedback form. The design difficulties including unmodelled dynamics and non-strict-feedback form are handled by resorting to a dynamic signal and the variable separation approach, respectively. A disturbance observer is constructed to cope with the effect of time varying disturbance. Neural networks are directly utilised to cope with the completely unknown non-linear functions and stochastic disturbances existing in systems. It is shown that the designed adaptive controller can guarantee that all the signals remain bounded and the desired signal can be tracked with a small domain of the origin. A numerical example is presented to illustrate the effectiveness of the proposed approach and an example of a real plant for one-link manipulator is provided to show the feasibility of the newly designed controller scheme.

Fixed-time adaptive neural tracking control for high-order nonlinear switched systems with input saturation and dead-zone

This paper is concerned with the problem of adaptive neural control for uncertain nonlinear strict-feedback time-delay systems with unknown virtual control coefficients. Radial basis function (RBF) neural networks are employed to directly approximate unknown virtual control signals, and then the adaptive neural control law is constructed by Lyapunov-Krasovskii functionals and backstepping. In order to avoid encountering a large number of adaptive parameters when using RBF neural networks as function approximators, an unknown constant, instead of unknown neural weights themselves, is employed as the estimated parameter. This technique makes only one adaptive parameter tuned online, thus significantly alleviating the burdensome computation. Meanwhile, some continuous functions are introduced to overcome the design difficulty originating from the use of one adaptive parameter. The proposed adaptive control guarantees the boundedness of all the signals in the closed-loop system. Simulation studies are presented to illustrate the effectiveness of the scheme.

Adaptive neural tracking control of pure-feedback nonlinear systems with unknown gain signs and unmodeled dynamics

Indirect adaptive neural tracking control of USVs under injection and deception attacks

A force/position hybrid adaptive control method based on radial basis function(RBF) neural network is proposed to solve the problem of difficulties to precisely describe the compliance and friction for robot terminal during trajectory tracking process.RBF neural network is an efficient feed-forward neural network with non-linear approximation and global optimization characteristics,which is not provided by other feed-forward networks,which is simple in network structure,and rapid in training speed.An adaptive controller is designed that relies on the nonlinear approximation ability of the RBF neural network to estimate the uncertainty factors in the models,the update rules for the weights of the controller neural network is provided and its finally uniform boundedness of the errors of the controller output force and position is proved.The controller is applied to a duct cleaning robot for simulation experiments.Simulation results shows that the adaptive controller demonstrates superior tracking precision and robustness compared with traditional adaptive controller.

This paper presents an adaptive neural tracking control approach for uncertain stochastic nonlinear time‐delay systems with input and output constraints. Firstly, the dynamic surface control (DSC) technique is incorporated into adaptive neural control framework to overcome the problem of ‘explosion of complexity’ in the control design. By employing a continuous differentiable asymmetric saturation model, the input constraint problem is solved. Secondly, the appropriate Lyapunov‐Krasovskii functional and the property of hyperbolic tangent functions are used to deal with the unknown time‐delay terms, RBF neural network is utilized to identify the unknown systems functions, and barrier Lyapunov functions (BLFs) are designed to avoid the violation of the output constraint. Finally, based on adaptive backstepping technique, an adaptive neural control method is proposed, and it decreases the number of learning parameters. Using Lyapunov stability theory, it is proved that the designed controller can ensure that all the signals in the closed‐loop system are 4‐Moment (or 2 Moment) semi‐globally uniformly ultimately bounded (SGUUB) and the tracking error converges to a small neighborhood of the origin. Two simulation examples are provided to further illustrate the effectiveness of the proposed approach.

For the social robot with serial elastic actuator, the joint dynamics model has the problems of strong coupling and high nonlinear, and the traditional PD control algorithm cannot achieve accurate trajectory tracking effect on the joint position of social robot using series elastic actuator. Therefore, an optimized Radial basis function (RBF) neural network adaptive control algorithm was proposed. The method based on RBF neural network approximates the social robot joint model parameters, an adaptive law was designed to estimate the weights of the neural network and the joint model online. The dynamic plane method is combined to improve the robustness of the control algorithm. The simulation results show that the trajectory tracking error peak of PD control algorithm is 0.2 rad. Compared with PD control algorithm, the trajectory tracking error peak of RBF neural network adaptive control algorithm based on dynamic surface optimization is reduced to ± 0.05 rad, which realizes accurate approximation of the parameters of social robot joint model, and accurate dynamics model approximation provides a theoretical basis for further research on human–robot interaction (HRI) of social robots.

This paper is dedicated to neural networks-based adaptive finite-time control design of switched nonlinear systems in the time-varying domain. More specifically, by employing the approximation ability of neural networks system, an integrated adaptive controller is constructed. The main aim is to make sure the closed-loop system in arbitrary switching signals is semi-global practical finite-time stable (SGPFS). A backstepping design with a common Lyapunov function is proposed. Unlike some existing control schemes with actuator failures, the key is dealing with the time-varying fault-tolerant job for the switched system. It is also proved that all signals in the system are bounded and the tracking error can converge in a small field of the origin in finite time. A practical example is presented to illustrate the validity of the theory.