Self-tuning adaptive control of multi-input, multi-output nonlinear systems using multilayer recurrent neural networks with application to synchronous power generators

In this paper, an adaptive neural networks control approach is proposed for a class of multi-input multi-output (MIMO) non-affine nonlinear dynamic systems in the presence of input saturation. The difficulty in controlling the saturated non-affine system is overcome by introducing a system transformation, so as the system can be reformulated as an affine of a canonical system. In the control design, neural networks are used in the online learning of the unknown dynamics and the input saturation is approximated to reduce the influence caused by the nonlinearities, and a robustifying control term is used to compensate for the approximation errors. Compared to the literature, in the proposed approach, the structure of the designed controller is much simpler since the causes for the problem of complexity growing in existing methods are eliminated. The stability analysis of the closed-loop system is investigated by using Lyapunov theory. Numerical simulation illustrated the proposed control scheme with satisfactory results.

This manuscript introduces a novel methodology to solve the state estimation of discrete-time multi-input multi-output (MIMO) nonlinear systems with uncertain dynamics. The mathematical model of the nonlinear systems considered in this paper satisfies the usual Lagrangian structure that characterizes many mechanical, electrical or electromechanical systems. A recurrent neural network (RNN) estimates the uncertain dynamics of the MIMO system with an updating law based on a particular variant of the discrete-time version of the super-twisting algorithm (DSTA). A Lyapunov stability analysis defines the convergence zone for the state estimation error throughout the solution of a matrix inequality. The convergence zone for the estimation is smaller when the DSTA and the RNN work together in an observer. Numerical examples demonstrate how the adaptive observer reduces the zone of convergence and the oscillations in the steady state compared with a discrete version of the STA with additional linear correcting terms. An experimental implementation shows how the observer estimates the unknown states of a Van Der Pol Oscillator. A comparison against some variations of the DSTA justifies the advantages of the mixed DSTA-RNN observer.

This paper provides an inverse-model-based iterative learning control (ILC) for the unknown multi-input multi-output (MIMO) nonlinear system with neural network (NN), where a novel gradient adaptive law is used to update the NN weights both hidden and output layers such a faster convergence can be achieved. First, a three-layer NN structure is introduced to observe the MIMO nonlinear system with input–output data, and a new gradient algorithm is proposed to update the unknown parameters of both hidden and output layers. Then, the input dynamic can be obtained with the NN observer, and the inversion-model-based control is designed. Moreover, the ideal inversion control can be obtained based on the reference signal, and the inverse ILC is designed. The stability of the NN observer and the convergence of the inverse-model-based control are analyzed. Finally, a SCARA manipulator MIMO model is simulated to illustrate the correctness of the proposed methods.

In this article, an improved model-free adaptive control (iMFAC) is proposed for discrete-time multi-input multioutput (MIMO) nonlinear systems with an event-triggered transmission scheme and quantization (ETQ). First, an event-triggered scheme is designed, and the structure of the uniform quantizer with an encoding-decoding mechanism is given. With the concept of partial form dynamic linearization based on event-triggered and quantization (PFDL-ETQ), a linearized data model of the MIMO nonlinear system is constructed. Then, an improved model-free adaptive controller with the ETQ process is designed. By this design, the update of the pseudo partitioned Jacobean matrix (PPJM) estimates and control inputs occurs only when the trigger conditions are met, which reduces the network transmission burden and saves the computing resources. Theoretical analysis shows that the proposed iMFAC with the ETQ process can achieve a bounded convergence of tracking error. Finally, a numerical simulation and a biaxial gantry motor contour tracking control system simulation are given to illustrate the feasibility of the proposed iMFAC method with the ETQ process.

This paper presents a method of continuous-time simple-adaptive control (SAC) for multi-input multi-output (MIMO) nonlinear systems using multifraction neural networks. The control input is given by the sum of the output of the simple adaptive controller and the output of the multifraction neural network is used to compensate the nonlinearity of plant dynamics that is not taken into consideration in the usual SAC. The role of the multifraction neural network is to construct a linearized model by minimizing the output error caused by nonlinearities in the control systems.

This article focuses on adaptive finite time control of multi-input multi-output (MIMO) nonlinear systems under time-varying actuator failures.A finite time adaptive control policy is proposed, guaranteeing the practical finite-time stability of the MIMO nonlinear systems with actuator failures. Simulation study illustrates the usefulness of the proposed policy.

The paper proposes an approach to second order sliding mode control for multi-input multi-output (MIMO) nonlinear uncertain systems. Second order sliding mode techniques for single-input single-output (SISO) systems show, with respect to standard sliding mode control, the same properties of robustness and precision, feature a higher order accuracy and can be exploited to eliminate the chattering effect. The application of second order sliding mode control to MIMO nonlinear systems is a challenging matter. The presented MIMO control method applies to a quite large class of nonlinear processes affected by uncertainties of general nature; the control design is simple; the conditions of existence on the controllers are weak. The proposed procedure represents a general approach to the second order sliding mode control of MIMO systems

We address the problem of global Lyapunov stability of discrete-time recurrent multilayer neural networks (RMLNN) in the unforced (unperturbed) setting. It is assumed that network weights are fixed to some values, for example, those attained after training. To apply the method of reduction of attractor estimate, we use the state space extension method to present RMLNN in the form of discrete-time dynamical system. We describe also a new algorithm for checking the global asymptotic stability of RMLNN, which is also based on the method of reduction of attractor estimate, and is much better from the computational viewpoint. An example shows the efficiency of this new algorithm.

A robust controller is developed for a class of multi-input multi-output (MIMO) uncertain nonlinear systems for which the robust backstepping method may not be applicable directly. The new design employs robust backstepping and forwarding construction of a Lyapunov function.

An optimal adaptive actuator failure compensation control scheme is proposed for a class of multi-input multi-output (MIMO) uncertain nonlinear continuous-time systems with redundant actuators subject to uncertain failures. Only single neural network is used in the proposed adaptive dynamic programming (ADP) scheme to simultaneously learn the infinite horizon cost function and the approximate optimal control input that minimizes the Hamilton-Jacobi-Bellman (HJB) equation. Besides, an uncertainty estimator is designed to estimate the uncertain dynamics and an adaptive parameter adjustment law is designed to estimate the actuator failure coefficients. Stability analysis and simulation results can demonstrate the effectiveness of the proposed method.

A multi-input multi-output nonlinear feedback system is considered in this paper by using operator based robust right coprime factorization approach. Robust stabilization and output tracking performance of the MIMO system are studied. That is, some conditions of the robust right coprime factorization for the MIMO nonlinear system are derived. Using the proposed conditions, the MIMO system can be controlled robustly and the desired output tracking performance is realized. Simulation example is given to support the theoretical analysis.

A method is presented for the generation of parameter sensitivity functions for multi-input multi-output nonlinear sampled-data systems on a hybrid or on a digital computer. An example is presented illustrating the approach.

This paper studies an adaptive tracking control for a class of multi-input multi-output (MIMO) non-affine nonlinear systems, with input dead-zone nonlinearity and external disturbances. By using the mean-value theorem, the system model is transformed into an affine form so as the difficulty in controlling non-affine systems is overcome. In the proposed control design, neural networks (NNs) are used to approximate the unknown nonlinearities based on their universal approximation properties. To compensate for approximation errors and external disturbances, an adaptive robust control term is introduced. In comparison with existing approaches, the structure of the designed controller is considerably simpler, and can handle a wider range of nonlinear systems. The stability of the closed-loop system is investigated by using Lyapunov theory. The simulation results illustrate the proposed method.

An advanced robust nonlinear control scheme is proposed for single-machine infinite-bus power system (SMIS) equipped with thyristor controlled series compensation (TCSC). Firstly, for a class of multi-input multi-output (MIMO) nonlinear system with disturbances, a disturbance attenuation method in the sense of L 2 gain is proposed through recursive method to construct storage function, and then disturbance attenuation control law is obtained. Secondly, in view of the strong nonlinearity and uncertainty of power system, a nonlinear model with disturbances is constructed. Lastly, a new robust nonlinear coordinated control law of TCSC and generator excitation is developed. Simulation results on SMIS show clearly that the proposed control law can attenuates the influence of external disturbances on the system and transient stability of power system is improved.

[Abstract] This paper presents the design of a higher-order sliding mode controller for finite-time stabilization of a multi-input multi-output nonlinear aeroelastic system with uncertain dynamics using leadingand trailing-edge control surfaces. The model describes the plunge and pitch dynamics of a prototypical wing section. It is assumed that the system parameters are not known precisely and external disturbance inputs are present in the model. The uncontrolled model exhibits limit cycle oscillations. The objective is to control the state vector of the system to the origin. A robust control law is designed for the finite-time trajectory control of the plunge and pitch angle trajectory. The control law includes a nominal finite-time stabilizing control signal designed for the system without uncertainties, and a discontinuous control signal for nullifying the effect of uncertain functions in the model. For the nominal control law design, the notion of geometric homogeneity is utilized and the discontinuous part of the controller is derived based on a second-order sliding mode control scheme. In the closed-loop system including the nominal and discontinuous control signals, finite-time control of the complete state vector associated with the plunge-pitch motion of the system to the origin is accomplished. Simulation results are presented which show the suppression of the limit cycle oscillations, despite parameter uncertainties and triangular, exponential, sinusoidal and random gust loads.