Neuro Controller Research Articles

Ascendancy of level in nonlinear tank system by neuro controller

The main motive of this paper is to control a nonlinear process tank system using an intelligent neuro-controller-based Artificial Neural Network (ANN). To reduce the mean squared error Proportional–Integral–Derivative (PID) and fuzzy controllers were evaluated with Bayesian Regularisation. Moreover, Bayesian regularization can also disclose potentially complex relationships, where it finds its application in quantitative research resulting in robust structure, thereby justifying its usage in the proposed system. Here the network training was based on the Levenberg–Marquardt algorithm. Two neural controllers were employed to emphasize the plant’s output response during step and nonlinear responses.The output of the neuro-controller was found to have a decreased overshoot time and undershoot time. The results of the proposed method’s settling time, overshoot, and undershoot were compared. The comparative results indicate a better performance of the proposed model. The MATLAB platform was used to complete the entire project.

Neuronskih mreža u upravljačkoj šemi sa unutrašnjim modelom

In this research, control of the Direct Current motor is accomplished using a neuro controller in the Internal Model Control scheme. Two Feed Forward Neural Networks are trained using historical input-output data. The first neural network is trained to identify the object's dynamic behavior, and that model is used as an internal model in the control scheme. The second neural network is trained to obtain an inverse model of the object, which is applied as a neuro controller. Experiment is conducted on the real direct current motor in laboratory conditions. Obtained results are compared to those achieved by implementing the Direct Inverse Control method with the same neuro controller. It was demonstrated that the proposed control method is simple to implement and the system robustness is achieved, which is a great benefit, aside from the fact that no mathematical model of the system is necessary to synthesize the controller of the real object.

Machine Learning Techniques for Increasing Efficiency of the Robot's Sensor and Control Information Processing.

Real-time systems are widely used in industry, including technological process control systems, industrial automation systems, SCADA systems, testing, and measuring equipment, and robotics. The efficiency of executing an intelligent robot’s mission in many cases depends on the properties of the robot’s sensor and control systems in providing the trajectory planning, recognition of the manipulated objects, adaptation of the desired clamping force of the gripper, obstacle avoidance, and so on. This paper provides an analysis of the approaches and methods for real-time sensor and control information processing with the application of machine learning, as well as successful cases of machine learning application in the synthesis of a robot’s sensor and control systems. Among the robotic systems under investigation are (a) adaptive robots with slip displacement sensors and fuzzy logic implementation for sensor data processing, (b) magnetically controlled mobile robots for moving on inclined and ceiling surfaces with neuro-fuzzy observers and neuro controllers, and (c) robots that are functioning in unknown environments with the prediction of the control system state using statistical learning theory. All obtained results concern the main elements of the two-component robotic system with the mobile robot and adaptive manipulation robot on a fixed base for executing complex missions in non-stationary or uncertain conditions. The design and software implementation stage involves the creation of a structural diagram and description of the selected technologies, training a neural network for recognition and classification of geometric objects, and software implementation of control system components. The Swift programming language is used for the control system design and the CreateML framework is used for creating a neural network. Among the main results are: (a) expanding the capabilities of the intelligent control system by increasing the number of classes for recognition from three (cube, cylinder, and sphere) to five (cube, cylinder, sphere, pyramid, and cone); (b) increasing the validation accuracy (to 100%) for recognition of five different classes using CreateML (YOLOv2 architecture); (c) increasing the training accuracy (to 98.02%) and testing accuracy (to 98.0%) for recognition of five different classes using Torch library (ResNet34 architecture) in less time and number of epochs compared with Create ML (YOLOv2 architecture); (d) increasing the training accuracy (to 99.75%) and testing accuracy (to 99.2%) for recognition of five different classes using Torch library (ResNet34 architecture) and fine-tuning technology; and (e) analyzing the effect of dataset size impact on recognition accuracy with ResNet34 architecture and fine-tuning technology. The results can help to choose efficient (a) design approaches for control robotic devices, (b) machine-learning methods for performing pattern recognition and classification, and (c) computer technologies for designing control systems and simulating robotic devices.

Design and implementation of hybrid energy sources with fuzzy neuro control for DC micro grid system used for electric vehicle

Design and implementation of hybrid energy sources with fuzzy neuro control for DC micro grid system used for electric vehicle

Design and implementation of hybrid energy sources with fuzzy neuro control for DC micro grid system used for electric vehicle

Design and implementation of hybrid energy sources with fuzzy neuro control for DC micro grid system used for electric vehicle

Adaptive Neural Networks Based Robust Output Feedback Control for Nonlinear System

The performance of the control system is reduced by uncertain nonlinearities behaviors, which can be enhanced by implementing an adaptive approach represented by the robust output-feedback control and artificial neural network, which is proposed in this paper and utilized for identification and control of a nonlinear system. The Cart Pole System (CPS) is treated as a multi-body dynamical system, and the nonlinear swing-up problem is handled by designing an adaptive neural network which trained using a modified conventional controller called Linear Quadratic Optimal State Estimator with Integral Control (LQOSEIC). In this paper, the nonlinear system CPS stabilized utilizing robust output feedback control called LQOSEIC, this controller allows a linearized system to act as a model reference for the original nonlinear system, but they are only valid for a limited range of operations and will fail if the plant characteristics are unknown or uncertainty. An adaptable neural network is used to overcome this challenge., in which the adaptive neuro controller is trained offline using LQOSEIC to get the initial weights of neurons for layers network, after finished the training the LQOSEIC will be replaced by adaptive neural control. The real advantage of a neuro-controller is its ability to update online depending on the error signal. The neuro-controller demonstrates that when any disturbance or uncertainty arises in a non-linear system, neural networks characterized by online learning compensate for the effect of unpredictable conditions. The suggested adaptive neural network improves control performance and ensures the closed-loop control system's robust stability. Finally, numerical simulations are used to demonstrate the efficacy of the proposed controllers.

Cooperative adaptive emotional neuro-control for a class of higher-ordered heterogeneous uncertain nonlinear multi-agent systems

Emotion is nature’s most exquisite survival mechanism to speedily and reasonably handle the uncertainties and complexities of life. It is hence reasonable to expect emotion-based artificial agencies to better handle their complex and uncertain environments and their interactions with other known or unknown agencies. Accordingly, an emotional neural network is proposed here for multi-agent systems (MAS). This computational paradigm has the universal approximation property of its well-established predecessor, the multilayer perceptron, but also promises the characteristics of the emotional mind such as fast response and learning. Specifically, we extend our previously established radial basis emotional neural network (RBENN) to approximate the uncertain dynamics in a cooperative adaptive radial basis emotional neuro-controller (CARENC) for a class of higher-ordered uncertain nonlinear MAS. Artificial potential functions are used to describe the interactions among heterogeneous agents. Also, the approximation error is compensated using an adaptive component. Suitable update laws are derived for the weights of RBENN that are consistent with the basic emotional models. In following Lyapunov stability theory, the overall system stability is also guaranteed. Application to two nonlinear multi-agent control problems shows a lower steady-state tracking and lesser control energy consumption when compared with a competing multi-agent neuro controller.

Adaptive Locomotion Control of a Hexapod Robot via Bio-Inspired Learning.

In this paper, an adaptive locomotion control approach for a hexapod robot is proposed. Inspired from biological neuro control systems, a 3D two-layer artificial center pattern generator (CPG) network is adopted to generate the locomotion of the robot. The first layer of the CPG is responsible for generating several basic locomotion patterns and the functional configuration of this layer is determined through kinematics analysis. The second layer of the CPG controls the limb behavior of the robot to adapt to environment change in a specific locomotion pattern. To enable the adaptability of the limb behavior controller, a reinforcement learning (RL)-based approach is employed to tune the CPG parameters. Owing to symmetrical structure of the robot, only two parameters need to be learned iteratively. Thus, the proposed approach can be used in practice. Finally, both simulations and experiments are conducted to verify the effectiveness of the proposed control approach.

Improving Wind Turbine Pitch Control by Effective Wind Neuro-Estimators

Data about wind are usually available from different databases, for different locations. In general, this information is the average of the wind speed over time. The wind reports are crucial for designing wind turbine controllers. But when working with floating offshore wind turbines (FOWT), two problems arise regarding the wind measurement. On the one hand, there are no buoys at deep sea, but near the coast where the wind is not so strong neither so stable; so the measurements do not fully correspond to reality. On the other hand, these floating devices are subjected to extreme environmental conditions (waves, currents, ...) that produce disturbances and thus may distort wind measurements. To address this problem, this work presents a novel pitch neuro-control architecture based on neuro-estimators of the effective wind. The control system is composed of a proportional-integral-derivative (PID) controller, a lookup table, a neuro-estimator, and a virtual sensor. The neuro-estimator is used to estimate the effective wind in the FOWT and to forecast its future value. Both current and future wind signals are combined and power the controller. The virtual sensor also provides a measure of the effective wind based on other available signals related to the wind turbine, such as the pitch angle and the angular velocity of the generator. Neural networks are trained online to adapt to changes in the environment. Intensive simulations are carried out to validate the effectiveness of this neuro control approach. Controller performance is compared to a PID, obtaining better results. Indeed, an improvement of 16% for sinusoidal wind and an average improvement of 8% are observed.

Neural network implementation of controllers for multi-channel objects synthesized by polynomial method

The implementation of neural network multichannel controllers synthesized by polynomial matrix decomposition is analysed. Objects and controllers are assumed to be linear; these allow them to be described by matrix transfer functions. The transfer function of the object is converted to the right polynomial matrix inter-simple decomposition. The transfer function of the controller is sought in the form of a left polynomial matrix of inter-simple decomposition that allows leading the characteristic matrix to the form of a linear matrix polynomial equation with two matrix indeterminates. This equation is solved by leading to a matrix equation with numeric matrix indeterminates. Then the controller equation is converted to a discrete equation. The discrete sampling step is chosen small enough to allow the systems with continuous and discrete controllers have sufficiently close transient processes. The discrete controller is converted to a structure including delay elements, adder units and amplification coefficients. Then this structure is presented in the form of a set of neurons. The operation of the algorithm is illustrated by the example of the synthesis of an unstable inverted pendulum control, which includes two PID controllers. Possible increases in neuro controller performance are demonstrated. Two PID controllers are combined into one neural network in order to further optimization.

Performance Evaluation of Effective Controller for Buck-Boost DC-DC DSNIBBC

This paper deals about the development of intelligent controllers (ICs) for load voltage regulation of a double switch noninverting buck-boost DSNIBBC operated in pseudo continuous coil current (PCCC). The DSNIBBC is a kind of DC-DC converter, which converts the positive DC source voltage into the positive DC load voltage (buck to boost mode and boost to buck mode). Due to the time changing, ON/OFF characteristics along with its transient and dynamic characteristics of DSNIBBC have become nonlinear. Since the classical linear controllers are inappropriate to enhance the transient and dynamic performances of DSNIBBC, ICs are designed to improve the load control of the same DSNIBBC operated in transient and dynamic regions. In this paper, two ICs, namely, fuzzy logic controller (FLC) and neuro controller (NC) incorporated in the load voltage regulation loop are utilized. The performance comparison between the ICs and the classical proportional integral (PI) controller is discussed and validated at various regions through both software and prototype field programmable gate array [FPGA] plants. The responses are depicted to support the designed ICs.

Performance Improvement of a DC/DC Converter Using Neural Network Controller in comparison with Different Controllers

Controlling a system is a complicated job, especially when we talk about the nonlinearity of the system introduced by the external changes. This paper presents the procedure of designing, analysis, and verification of nonlinear autoregressive moving average controller (NARMA L2) as an artificial intelligence technique to track the output voltage of a Buck dc/dc converter in comparison with PID controller, digitalized sliding mode controller so as to reduce the ripples in output voltage and to suppress the transient overshoots, or in other words, enhance the transient response diversity of the plant in the case of load and line changes. In this technique, a back-propagation learning algorithm is derived to increase the effectiveness of the proposed controller. Finally, the proposed method of control using a neural network controller is designed using MATLAB/SIMULINK and the results of the converter for the Neuro controller are compared with different techniques of control.

Design and Implementation of Intelligent Hybrid Neuro Controller for Automatic Voltage Regulator

In this paper, we presented a hybrid controller by combining the advantages of both PID and Neuro controllers for automatic voltage regulator (AVR) system. A Neuro controller is designed using multilayer feedforward neural network and Levenberg-Marquardt backpropagation algorithm is used for training the network. Also, hybrid controller is achieved by blending the characteristics of classical PID and proposed Neuro controller using a switching mechanism based on the error. The proposed PID, Neuro and Hybrid controllers are simulated in MATLAB environment and their transient response parameters are compared. The simulation results clearly indicated the improvement in the transient output of the automatic voltage regulator system with proposed hybrid controller even in the presence of uncertainties in the system.

Neuro control of electric drive with variable moment of inertia

Neuro control of electric drive with variable moment of inertia

Improving the Intelligent Control of Magnetic Levitiation Ball Using Artificial Neural Network

Forecasting the time series behavior of magnetic levitation system has been a major research objective for the last five decades. This is due to the challenges presented as a result of its dynamic nature in motion. Initially this problem was supposed to be solves by control engineering researchers using predictive control modeling. However, due to the robust nature of the technological application (high speed train, maglev ball, e.t.c), precision in the control system has been a major challenge according to Earnshaw’s theorem. This paper improves intelligent control of a magnetic levitation (ball) system using artificial neural network. First the maglev system plant is identified and then a Neuro controller is designed to predict the dynamic behavior of this system using feedback linearization process. The system will be design and simulated using neural network tool and Mathlab.

Response Simulation, Data Cleansing and Restoration of Dynamic and Static Measurements Based on Deep Learning Algorithms

In this study, an output-based neuro controller was built based on the idea of the adaptive neuro-fuzzy inference system (ANFIS) and its capabilities in response simulation, data cleansing and restoration capability were verified using measurement data from actual structural testing. The ANFIS is a family of the deep learning algorithm, which incorporates the benefits of adaptive control technique, artificial neural network, and the fuzzy inference system. Thus, it is expected to produce very accurate predictions even for the highly nonlinear system. Forced vibration responses of a five-story steel building were simulated by ANFIS and its accuracy was compared with the results of Recurrent Neural Network (RNN), which is a type of traditional artificial neural networks. Simulations by ANFIS were very accurate with a much lower root means square error (RMSE) than RNN. Simulated data by ANFIS showed an almost perfect match with the original. Even the small ripples in the power spectrum plot outside the dominant frequency were successfully reproduced. In addition, the ANFIS was used to increase the sampling rate of dynamic data. It was shown that missing high-frequency contents could be successfully reproduced when the ANFIS was properly trained. Lastly, The ANFIS was applied to remove the noise in the measured data from RC column cyclic load tests. The outliers were corrected effectively, but the tendency of flattening the peak values was observed.

Stable robust adaptive radial basis emotional neurocontrol for a class of uncertain nonlinear systems

Abstract Various models of emotional learning in mammalian brain have inspired a number of cognitive control and decision-making paradigms to date. These architectures are promising in terms of their learning ability and low computation, but they generally lack rigorous analysis of stability and mathematical proof of performance. In order to remedy this weakness, this paper proposes a radial basis emotional neural network (RBENN) structure using a modified Thalamus. Lyapunov analysis is then used to derive laws of adaptation within the context of an indirect adaptive robust control architecture. Accordingly, the stability of the proposed adaptive robust radial basis emotional neuro controller (ARBENC) is proved for a class of uncertain nonlinear systems. The proposed architecture is then applied to two nonlinear systems with partially unknown dynamics with or without disturbances and noisy conditions. In comparison with competing radial basis function neural, fuzzy, and emotional controllers, ARBENC exhibits fast learning while keeping a simple structure, leading to improved error performance and less overall computational time. These numeric results are consistent with earlier reports on emotion-based approaches, the main advantage here is ARBENC's amenity to theoretical analysis and proof of stability.

피부문신페인팅 전극과 뇌-컴퓨터 웨어러블 회로에 의한 영상 기억 문자 인식

Brain computer interface neuro diagnostics were described using an art painted tattoo membrane (PTM) by in vitro electrode on the skin surface. The PTM membrane of skin film was made from Carboxymethyl cellulose plant pulp back bone structure and mineral oil mixing three electrode systems. The analytical chrono cyclic potential windows were of 2.0 ~-2.0 V. The accumulation scan direction of oxidation and reduction current were used, and a sophisticated method of stripping accumulation was performed to - 2.0 V deposition potential. Furthermore, a 30 second stripping time and other para strengths were optimized, such as square wave amplification, increment potential, redox direction and differential delta strength. These were examined in an electrolyte solution, and a standard sensing of working range was attained to ㎎/L metal detection. A trace detection limit of optimum para meters can be applied to neuro assay of in vivo diagnostics for memory sensing and imagination character recognition. Developed here are the results demonstrating the usability for brain computer interfaced neuro control.

Neuro control – using neuro monitoring data to control networked active instruments

Abstract Intraoperative neurophysiological monitoring (IONM) can be used to monitor the neural integrity and help the surgeon to reduce iatrogenic patient damage. We present a networked digital assistance system based on real-time IONM data. The system automatically reduces the power of active medical ablation instruments when they endanger neural structures. We integrated a nerve monitor into a real-time network of medical devices. In doing so, neurophysiological signals are available to all connected medical devices in real-time. With the IONM data a so-called function module Neuro Control calculates and controls the maximum allowed power for active instruments like an ultrasonic dissector.

Adaptive near-optimal neuro controller for continuous-time nonaffine nonlinear systems with constrained input

In this paper, an identifiercritic structure is introduced to find an online near-optimal controller for continuous-time nonaffine nonlinear systems having saturated control signal. By employing tw...