Neural Networks For Adaptive Control Research Articles

Artificial neural networks for adaptive control of profiled haemodialysis in patients with renal insufficiency

Background and objectiveCurrently, haemodialysis treatment is performed using an open-loop control approach, with initial settings of parameters such as ultrafiltration rate and dialyser composition being adapted to the current haemodynamic condition of each patient, although unexpected events may require additional adjustments to be made.Therefore, an artificial neural network-based approach has been presented to automatically control the ultrafiltration rate according to the specific patient conditions during the haemodialysis session, in order to regulate body weight loss, and the elimination of electrolytes and uremic toxins. MethodsThis modelling task is performed using a mathematical model of fluid and solute exchange based on first principles, which is used to simulate the process of a haemodialysis session in a specific patient under SIMULINK in order to define the underlying dynamic equations. Alongside this, MATLAB neural network tools are used to adjust the settings of the automatic controller for different body weight loss regulation profiles and variable dialysate sodium conditions during haemodialysis treatment. ResultsComputer simulation results show the adequate performance of the body weight loss neuroadaptive control system when submitted to different haemodialysis patterns, uremic toxins and sodium elimination evolution under changing dialysate sodium conditions. ConclusionsThe proposed approach proves to be a valuable tool as a test bench for the assessment of alternate haemodialysis profiles aimed to improve the treatment of patients by preventing dialysis-induced haemodynamic complications. The adaptive nature of the model-based control approach here presented makes it possible to personalize the haemodialysis treatment to the specific needs of each patient.

Artificial Neural Networks for Adaptive Control of Profiled Haemodialysis in Patients with Renal Insufficiency

Artificial Neural Networks for Adaptive Control of Profiled Haemodialysis in Patients with Renal Insufficiency

EVALUATING THE POSSIBILITIES OF APPLYING AN ARTIFICIAL NEURAL NETWORK FOR CONTROL AND DIAGNOSTICS OF THE ELECTRIC DRIVE SYSTEMS

The paper investigates and evaluates the possibilities of using an artificial neural network. Attention is paid to the structural and operational features of the artificial neural networks, the learning processes used in them and their capabilities. The methods of application of artificial neural network for the purpose of control and diagnostics of dynamic systems of the electric drives are considered. A comparative analysis of the electric drive systems with PID regulators and neuroregulators was conducted. The expediency and necessity of improving the abilities of training artificial neural networks for adaptive control and diagnostics of the electric drives with incomplete description have been revealed, including the drives operating under random influences and dynamically changing modes. The main circumstances preventing the use of artificial neural networks, the laws of the choice of types and methods of optimization in the process of artificial neural networks (ANN) training and the lack of criteria for choosing the number of the neurons in the network are given. A review of well-known works devoted to the use of ANN in the electric drive systems, as well as a comparative analysis of the feasibility of using gradient and genetic methods of their training are carried out. The comparative analysis was carried out by summarizing the conclusions in various published works. Analysis shows that, in most cases, networks trained by genetic algorithms provide more accurate results, easier learning, and shorter duration. At the same time, in some cases, the use of the back propagation algorithm in the certain problems leads to better results. Thus, it can be stated that the use of the preferred algorithm depends on the formulation of the task.

Application of Adaptive Controller Neural Network Based on RBF NN for Temperature Control Electrical Resistance Furnace

Application of Adaptive Controller Neural Network Based on RBF NN for Temperature Control Electrical Resistance Furnace

МОДЕЛЬ АДАПТИВНОЙ СИСТЕМЫ УПРАВЛЕНИЯ ПОТОКОМ РЕАКТИВНОЙ МОЩНОСТИ НА ГРАНИЦЕ БАЛАНСОВОЙ ПРИНАДЛЕЖНОСТИ ПРЕДПРИЯТИЯ И СЕТЕВОЙ ОРГАНИЗАЦИИ

The paper proposes the structure of the reactive power flow control system at a balance affiliation boundary of an enterprise and a grid operator. It deals with the features of neural networks and presents a numerical mathematical model based on an artificial neural network for adaptive control of reactive power flow. Training of the neural network is automated and implemented in software. Levenberg-Marquardt method is used for neural network training. The structure of the control system of reactive power flow and voltage levels is designed. A Levenberg-Marquardt method-based algorithm for neural network training is developed and implemented in software. The authors have proposed an adaptive control system of reactive power flow at the balance affiliation boundary of an enterprise and a grid operator based on the synthesis of the artificial neural network and STATCOM own logic.

Lyapunov-Based Dynamic Neural Network for Adaptive Control of Complex Systems

In this paper, an adaptive neuro-control structure for complex dynamic system is proposed. A recurrent Neural Network is trained-off-line to learn the inverse dynamics of the system from the observation of the input-output data. The direct adaptive approach is performed after the training process is achieved. A lyapunov-Base training algorithm is proposed and used to adjust on-line the network weights so that the neural model output follows the desired one. The simulation results obtained verify the effectiveness of the proposed control method.

Application of Neural Network on Flight Control

Over the last three decades, adaptive control has evolved as a powerful methodology for designing feedback controller of nonlinear systems. Most of the studies assume that the system nonlinearities are known a prior, which is generally not applicable in the real world. To overcome this drawback, from twenty years ago, there has been a tremendous amount of activity in applying Neural Networks for adaptive control. With their powerful ability to approximate nonlinear functions, neuro-controllers can implement the expected objectives by canceling or learning the unknown nonlinearities of the system to be cancelled. Neural Networks are specially suitable for the adaptive flight control applications where system dynamics are dominated by the unknown nonlinearities. In the past decades, major advances have been made in adaptive control of linear time-invariant plants with unknown parameters. The choice of the controller structure is based on well established results in linear systems theory and stable adaptive laws which assure the global stability of the overall system are derived based on properties of those system. In recent years, Artificial Neural Network based control strategies have attracted much attention because of their powerful ability to approximate continuous nonlinear functions .(1) In fact, a neural controller with on-line learning can adapt to the change in a system dynamics and hence is an ideal choice for controlling highly nonlinear system with uncertainty. For adaptive control purposes neural networks are used as approximation models of unknown nonlinearities. The input / output response of neural network models is modified by adjusting the values of its adjusting the values of its parameters. Although it is true that polynomials, trigonometric series and orthogonal functions can also be used as function approximator, neural networks have been found to be particularly useful for controlling highly uncertain, nonlinear and complex systems. Neural control strategies can be broadly classified into off-line and on-line schemes based on how the parameters of the network are tuned. When the neural controller operates in an on-line mode, it has no a priori knowledge of the system to be controlled and the parameters of the network are updated while the input - output data is received. In the off-line control, the network's parameters are determined from the unknown training pairs and then those parameters are fixed for control purposes. (2)

Variable structure neural networks for adaptive control of nonlinear systems using the stochastic approximation

This paper is concerned with the adaptive control of continuous-time nonlinear dynamical systems using neural networks. Referred to as a variable structure neural network, a novel neural network architecture, is proposed and shown to be useful in approximating the unknown nonlinearities of dynamical systems. In the variable structure neural network, the number of basis functions can be either increased or decreased with time according specified design strategies so that the network will not overfit or underfit the data set. Based on the Gaussian radial basis function (GBRF) variable neural network, an adaptive control scheme is presented. The location of the centers and the determination of the widths of the GBRFs are analysed using a new method inspired from the adaptive diffuse element method combined with a pruning algorithm. In the standard problem of a feedback based control, the cost to be minimized is a function of the output derivative. When the cost function depends on the output error, the gradient method cannot be applied to adjust the neural network parameters. In this case, the stochastic approximation approach allows the computation of the cost function derivatives. The developed weight adaptive laws use a stochastic approximation algorithm. This algorithm consists of the use of the Kiefer–Wolfowitz method.

Variable neural networks for adaptive control of nonlinear systems

This paper is concerned with the adaptive control of continuous-time nonlinear dynamical systems using neural networks. A novel neural network architecture, referred to as a variable neural network, is proposed and shown to be useful in approximating the unknown nonlinearities of dynamical systems. In the variable neural networks, the number of basis functions can be either increased or decreased with time, according to specified design strategies, so that the network will not overfit or underfit the data set. Based on the Gaussian radial basis function (GRBF) variable neural network, an adaptive control scheme is presented. The location of the centers and the determination of the widths of the GRBFs in the variable neural network are analyzed to make a compromise between orthogonality and smoothness. The weight-adaptive laws developed using the Lyapunov synthesis approach guarantee the stability of the overall control scheme, even in the presence of modeling error(s). The tracking errors converge to the required accuracy through the adaptive control algorithm derived by combining the variable neural network and Lyapunov synthesis techniques. The operation of an adaptive control scheme using the variable neural network is demonstrated using two simulated examples.

Hybrid learning concepts based on self-organizing neural networks for adaptive control of walking machines

Online learning of complex control behaviour of autonomous mobile robots like walking machines is one of the current research topics. In this article, a hybrid learning architecture based on reinforcement learning (RL) and self-organizing neural networks for online adaptivity is presented. The hybrid concept integrates different learning methods and task-oriented representations as well as available domain knowledge. The proposed concept is used for RL of control strategies on different control levels on a walking machine.

A reinforcement learning neural network for adaptive control of Markov chains

In this paper we consider the problem of reinforcement learning in a dynamically changing environment. In this context, we study the problem of adaptive control of finite-state Markov chains with a finite number of controls. The transition and payoff structures are unknown. The objective is to find an optimal policy which maximizes the expected total discounted payoff over the infinite horizon. A stochastic neural network model is suggested for the controller. The parameters of the neural net, which determine a random control strategy, are updated at each instant using a simple learning scheme. This learning scheme involves estimation of some relevant parameters using an adaptive critic. It is proved that the controller asymptotically chooses an optimal action in each state of the Markov chain with a high probability.

Self-organising neural networks for adaptive control

Self-organizing neural networks have been implemented in a wide range of application areas such as speech processing, image processing, optimization and robotics. Recent variations to the basic model proposed by the authors enable it to order state space using a subset of the input vector and to apply a local adaptation procedure that does not rely on a predefined test duration limit. Both these variations have been incorporated into a new feature map architecture that forms an integral part of an Hybrid Learning System (HLS) based on a genetic-based classifier system. Problems are represented within HLS as objects characterized by environmental features. Objects controlled by the system have preset targets set against a subset of their features. The system's objective is to achieve these targets by evolving a behavioural repertoire that efficiently explores and exploits the problem environment. Feature maps encode two types of knowledge within HLS — long-term memory traces of useful regularities within the environment and the classifier performance data calibrated against an object's feature states and targets. Self-organization of these networks constitutes non-genetic-based (experience-driven) learning within HLS.

A Class of Recurrent Neural Networks for Adaptive Control of Nonlinear Systems

In this paper, recurrent neural networks (RNN) are used to construct nonlinear learning systems for a class of unknown nonlinear systems. We describe the RNN controllers with reference to adaptive control theory and put emphasis on their flexible structure to express nonlinear systems. The proposed control scheme employs two types of RNN: The first network learns the unknown dynamics of the nonlinear process to be controlled. The second utilises the information given by the first network about the process, to learn the inverse dynamics of the neural network process model such that the error between the system and the desired output is nullified. We also present a generalized criterion for training the RNN which gives rise to new control schemes. The training methodology that is used for the weights adaptation is much faster than classical backpropagation algorithm. We investigate the effects the type of nonlinear function, present in the generalized criterion, have on performance of the learning algorithm of unknown nonlinear system dynamics.

Neural network architecture for control

Two important computational features of neural networks are associative storage and retrieval of knowledge, and uniform rate of convergence of network dynamics independent of network dimension. It is indicated how these properties can be used for adaptive control through the use of neural network computation algorithms, and resulting computational advantages are outlined. The neuromorphic control approach is compared to model reference adaptive control on a specific example. It is shown that the utilization of neural networks for adaptive control offers definite speed advantages over traditional approaches for very-large-scale systems.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>