Adaptive Neural Network Method Research Articles

Neuro-Based Consensus Seeking for Nonlinear Uncertainty Multi-Agent Systems Constrained by Dead-Zone Input

The topic about consensus target track seeking for high-order nonlinear multi-agent systems (MASs) with unmodeled dynamics, dynamic disturbances, and dead-zone input is considered in the paper. Using the strong nonlinear map characteristic of radial basis function neural networks (RBFNNs), the complex functions arising from recursive procedure are simplified. Also, inspired by input-to-state practical stability (ISpS), the authors construct a dynamical signal in order to counteract the impact of unmodeled dynamics and dynamic disturbances. The bounded inequality expression has been applied to tackle the unknown input of dead zone. Based on this, consensus control protocol suitable for nonlinear constraints has been constructed by using the recursive backstepping technique and adaptive neural network method. Theoretical analysis indicates not only the uniform boundary of all signals in the closed-loop under the neuro-based consensus controller, but uniform ultimate convergence of consensus tracking errors. The final simulations also confirmed the correctness of the theoretical analysis.

Artificial neural network-based predictive model for supersonic ejector in refrigeration system

In this paper, a novel predictive model is proposed for the supersonic ejector in the refrigeration system based on the artificial neural network technique. A composite performance prediction framework for the supersonic ejector is developed by employing a back propagation neural network with particle swarm optimization (PSO-BPNN) algorithm and a dynamic error compensation method. Firstly, we study the model construction problem for the supersonic ejector based on the traditional BPNN modeling method and the PSO algorithm. And then, a dynamic compensation technique is given to improve the predictive accuracy of the ejector model based on the adaptive neural network method. Finally, both the simulation and experimental results are given to verify our method. The experimental validation results show that most prediction errors of our model are less than 4%. Therefore, compared with the existing traditional models, our model has a higher accuracy.

Adaptive Leader-Follower Consensus Control of Multiple Flexible Manipulators With Actuator Failures and Parameter Uncertainties

In this paper, the leader-follower consensus problem for a multiple flexible manipulator network with actuator failures, parameter uncertainties, and unknown time-varying boundary disturbances is addressed. The purpose of this study is to develop distributed controllers utilizing local interactive protocols that not only suppress the vibration of each flexible manipulator but also achieve consensus on joint angle position between actual followers and the virtual leader. Following the accomplishment of the reconstruction of the fault terms and parameter uncertainties, the adaptive neural network method and parameter estimation technique are employed to compensate for unknown items and bounded disturbances. Furthermore, the Lyapunov stability theory is used to demonstrate that followers' angle consensus errors and vibration deflections in closed-loop systems are uniformly ultimately bounded. Finally, the numerical simulation results confirm the efficacy of the proposed controllers.

Smooth-Switching Gain Based Adaptive Neural Network Control of n-Joint Manipulator with Multiple Constraints

Modeling errors, external loads and output constraints will affect the tracking control of the n-joint manipulator driven by the permanent magnet synchronous motor. To solve the above problems, the smooth-switching for backstepping gain control strategy based on the Barrier Lyapunov Function and adaptive neural network (BLF-ANBG) is proposed. First, the adaptive neural network method is established to approximate modeling errors, unknown loads and unenforced inputs. Then, the gain functions based on the error and error rate of change are designed, respectively. The two gain functions can respectively provide faster response speed and better tracking stability. The smooth-switching for backstepping gain strategy based on the Barrier Lyapunov Function is proposed to combine the advantages of both gain functions. According to the above strategy, the BLF-ANBG strategy is proposed, which not only solves the influence of multiple constraints, unknown loads and modeling errors, but also enables the manipulator system to have better dynamic and steady-state performances at the same time. Finally, the proposed controller is applied to a 2-DOF manipulator and compared with other commonly used methods. The simulation results show that the BLF-ANBG strategy has good tracking performance under multiple constraints and model errors.

Neural Network Command Filtered Control of Fractional-Order Chaotic Systems.

An adaptive neural network (NN) backstepping control method based on command filtering is proposed for a class of fractional-order chaotic systems (FOCSs) in this paper. In order to solve the problem of the item explosion in the classical backstepping method, a command filter method is adopted and the error compensation mechanism is introduced to overcome the shortcomings of the dynamic surface method. Moreover, an adaptive neural network method for unknown FOCSs is proposed. Compared with the existing control methods, the advantage of the proposed control method is that the design of the compensation signals eliminates the filtering errors, which makes the control effect of the actual system improve well. Finally, two examples are given to prove the effectiveness and potential of the proposed method.

A Gait Simulation and Evaluation System for Hip Disarticulation Prostheses

A little volume of lower extremity amputees needs hip disarticulation prostheses (HDPs). This situation makes it hard to invite hip amputees for prostheses test. The traditional evaluations for HDPs mostly refer to the questionnaires from the amputees and lack of quantitative criteria. In this article, we proposed a gait simulation and evaluation (GSE) system for HDPs testing. The GSE system contained a prosthetic thigh simulation robot that could reproduce the trajectory of human movement. We trained the simulation robot to walk with the tested HDPs using an adaptive neural network (ANN) control method. The ANN method was proposed to solve the unstable motion and large trajectory tracking error when the simulation robot was disturbed by the tested HDPs. The gait kinematic and kinetic data were collected to analyze the performance of the tested HDPs. The GSE system provided a stable test environment that opens up new possibilities for the HDPs test under controlled conditions. Note to Practitioners-Most of the prosthetic tests are based on amputees, especially hip disarticulation prostheses (HDPs). This work provides a gait simulation and evaluation system based on the quantified evaluation, which can repeatedly test the prostheses. The testing system improves the efficiency of the development of HDPs and decreases the security of the test with hip amputees. Furthermore, the system has backward compatibility for other lower limb prostheses (prosthetic knee and ankle).

SGD-Based Adaptive NN Control Design for Uncertain Nonlinear Systems.

In this paper, a stochastic gradient descent (SGD)-based adaptive neural network (NN) control scheme is presented for a class of uncertain nonlinear systems. The introduction of the SGD algorithm results in a better tracking performance compared with some other adaptive NN methods without using SGD. This is because the proposed SGD-based adaptive NN control strategy provides optimization algorithms for the weights, the widths, and the centers of the NNs, which can achieve a good function approximation performance. In order to implement the proposed method, extended differentiators are introduced to get the differential estimations of error signals, such that the loss function of the optimization algorithm can be constructed approximatively. Moreover, adaptive laws are designed to reduce the overall approximation errors, such that the tracking performance is further improved. By using the Lyapunov stability theory, it can be proved that the target signal is tracked by the system output within a small error. Finally, simulation and comparison results are given to show the effectiveness and advantages of the proposed method.

Safety-Guaranteed Trajectory Tracking Control for the Underactuated Hovercraft with State and Input Constraints

This paper develops a safety-guaranteed trajectory tracking controller for hovercraft by using a safety-guaranteed auxiliary dynamic system, an integral sliding mode control, and an adaptive neural network method. The safety-guaranteed auxiliary dynamic system is designed to implement system state and input constraints. By considering the relationship of velocity and resistance hump, the velocity of hovercraft is constrained to eliminate the effect of resistance hump and obtain better stability. And the safety limit of drift angle is well performed to guarantee the light safe maneuvers of hovercraft tracking with high velocities. In view of the natural capabilities of actuators, the control input is constrained. High nonlinearity and model uncertainties of hovercraft are approximated by employing adaptive radical basis function neural networks. The proposed controller guarantees the boundedness of all the closed-loop signals. Specifically, the tracking errors are uniformly ultimately bounded. Numerical simulations are implemented to demonstrate the efficacy of the designed controller.

Tracking control of a planar five-link bipedal walking system with point contact, considering self-impact joint constraint by adaptive neural network method

In order to achieve the practical characteristics of natural bipedal walking, a key feature is to realize "the straight knee state of walking" during stance and swing motions. Considering a straight knee necessitates that the shank link of each leg not to undergo the rotation angles which are greater than that of the thigh link. For this purpose, various methods have been proposed; the joint self-impact constraint has been suggested for energy-efficient (natural) bipedal walking while realizing the straight knee constraint. The prominent objective of this research is to present a model based control method for trajectory tracking of a normal human-like bipedal walking, by considering the joint self-impact constraint. To achieve this objective, the dynamical equations of motion of an unconstrained biped are taken, developed and then modified to consider the joint self-impact constraint at the knee joint. To control this complicated dynamical system, the available anthropometric normal gait cycle data are taken to generate the desired trajectories of the thigh and knee joints of the self-impact biped. Due to the existence of complex nonlinear terms in the dynamical governing equations of self-impact biped, the authors propose to design a nonlinear intelligent controller by taking advantage of the adaptive neural network control method, which neither requires the evaluation of inverse dynamical model nor the time consuming training process. According to the simulation results, the tracking control of the biped robot is accomplished well and the biped walking seems naturally, despite of involving complex nonlinear terms in the dynamical governing equations of the self-impact biped.

Control of a Ball on Sphere System with Adaptive Neural Network Method for Regulation Purpose

Control of a Ball on Sphere System with Adaptive Neural Network Method for Regulation Purpose

Performance enhancement of respiratory tumor motion prediction using adaptive support vector regression: Comparison with adaptive neural network method

ABSTRACTThe breathing motion moves internal organs and targeted regions determined by radiation therapy planning. For the radiation therapy, accurate prediction for breathing motion is of great interest as the outer targeted region treatment could endanger sensitive tissue. In this study, the use of a prediction algorithm with adaptive support vector regression (aSVR) was proposed and compared with the adaptive neural network (ANN) algorithm considering the prediction accuracy and training and predicting time. Respiration data from 87 patients treated by radiation therapy, were acquired with an optical marker at 30 Hz. Five types of prediction filters with the ANN or aSVR filters, were implemented and their performance was compared according to the size of the sliding window (2.5 and 5.0 sec), and the prediction latencies (100, 200, 300, 400, and 500 msec). Training and testing of the prediction algorithms with aSVR and ANN were performed. The root mean square error (RMSE) was used as the accuracy metric. The aSVR with an RBF kernel outperformed other prediction filters, including not only various types of ANN filters but also the aSVR with a linear kernel. A sliding window of 2.5 sec significantly and independently enhanced the overall accuracy. Otherwise, the training and prediction testing times were significantly prolonged in case of aSVR with an RBF kernel. The aSVR filter with the RBF kernel is in all cases superior to other filters regarding its accuracy; it also shows clinically applicable results from the viewpoint of training and predicting time, which may be effective for predicting patient breathing motion and thus enhancing the efficacy of radiation therapy.

Long-Range Prediction for Real-Time MPEG Video Traffic: An $H_{\infty}$ Filter Approach

A novel prediction scheme is proposed for real-time MPEG video to predict the burst and long-range dependent traffic. The trend and periodic characteristics of MPEG video traffic are fully captured by a proposed stochastic state-space dynamic model. Then a recursive Hinfin filtering algorithm is proposed to estimate traffic for long-range prediction. Simulation results based on real MPEG traffic data show that the proposed scheme has superior performance and lower complexity than some adaptive neural network methods, such as TDNN, NARX, and Elman neural networks.

Adaptive RBF network for parameter estimation and stable air–fuel ratio control

In the application of variable structure control to engine air-fuel ratio, the ratio is subjected to chattering due to system uncertainty, such as unknown parameters or time varying dynamics. This paper proposes an adaptive neural network method to estimate two immeasurable physical parameters on-line and to compensate for the model uncertainty and engine time varying dynamics, so that the chattering is substantially reduced and the air-fuel ratio is regulated within the desired range of the stoichiometric value. The adaptive law of the neural network is derived using the Lyapunov method, so that the stability of the whole system and the convergence of the networks are guaranteed. Computer simulations based on a mean value engine model demonstrate the effectiveness of the technique.

Adaptive neural network algorithms for computing proximity effect corrections

Computing self-consistent local dose corrections for images to offset proximity effects poses significant problems because the number of computations needed cannot be done in a reasonable amount of time. Recently, we have demonstrated the use of an adaptive neural network method to increase the speed of these computations by several orders of magnitude. We have now implemented these corrections in practical hardware, and have introduced improvements in the algorithm that further reduce the computational complexity and time. The challenges in computing proximity effect corrections are to find algorithms that work well for general feature shapes, and that can be efficiently implemented. This paper will discuss the iterative computation of optimal corrections for electron scattering, and the limits of image resolution that can be obtained. It will describe the neural network algorithm used to obtain equivalent results more efficiently, and the method of adaptively determining the network’s parameters. Finally, it will discuss several recent modifications that significantly improve the network’s performance.