Recurrent Fuzzy Neural Network Research Articles

INS/GNSS integration using recurrent fuzzy wavelet neural networks

In recent years, aided navigation systems through combining inertial navigation system (INS) with global navigation satellite system (GNSS) have been widely applied to enhance the position, velocity, and attitude information of autonomous vehicles. In order to gain the accuracy of the aided INS/GNSS in GNSS gap intervals, a heuristic neural network structure based on the recurrent fuzzy wavelet neural network (RFWNN) is applicable for INS velocity and position error compensation purpose. During frequent access to GNSS data, the RFWNN should be trained as a highly precise prediction model equipped with the Kalman filter algorithm. Therefore, the INS velocity and position error data are obtainable along with the lost intervals of GNSS signals. For performance assessment of the proposed RFWNN-aided INS/GNSS, real flight test data of a small commercial unmanned aerial vehicle (UAV) were conducted. A comparison of test results shows that the proposed NN algorithm could efficiently provide high-accuracy corrections on the INS velocity and position information during GNSS outages.

Smart backstepping control using revised recurrent fuzzy neural network and revised ant colony optimization for linear permanent magnet synchronous motor drive system

Because of the uncertainty’s action in a linear permanent magnet synchronous motor drive system such as the external load force, the cogging force, the column friction force and the Stribeck effect force and the parameters variations, it is difficult to reach specific control performances by using the existing linear controller. To raise robustness under occurrence of parameters uncertainties and external force disturbances, the smart backstepping control system with three adaptive laws is proposed for controlling the linear permanent magnet synchronous motor drive system. In accordance with the Lyapunov function, three adaptive laws are derived to ameliorate the system’s robustness. Furthermore, the smart backstepping control system using revised recurrent fuzzy neural network and revised ant colony optimization with the compensated controller is proposed to improve the control performance. The revised recurrent fuzzy neural network acts as the estimator of the uncertainty’s disturbances. In addition, the compensated controller with error estimation law is proposed to compensate the minimum rebuilt error. Moreover, two learning rates of the weights in the revised recurrent fuzzy neural network are derived according to the discrete-type Lyapunov stability to assure convergence of the output tracking error and are adopted by using the revised ant colony optimization to speed-up parameter’s convergence. Finally, some comparative performances are verified through some tentative upshots that the smart backstepping control system by virtue of revised recurrent fuzzy neural network and revised ant colony optimization with the compensated controller results in better control performances for the linear permanent magnet synchronous motor drive system.

RETRACTED ARTICLE: Integral Backstepping Control of LPMSM Drive System Using Revised Recurrent Fuzzy NN and Mended Particle Swarm Optimization

A linear permanent magnet synchronous motor (LPMSM) drive system is keeping in many nonlinear effects such as the external load force, the flux saturation, the cogging force, the column friction and Stribeck force, and the parameters variations. Due to the uncertainty effects the existing linear controllers can not achieve better control performances for the LPMSM drive system. To raise robustness under occurrence of uncertainty, the integral backstepping control system with hitting function is proposed for controlling the LPMSM drive system in accordance with the Lyapunov function. To improve larger chattering phenomenon under uncertainties effects, the integral backstepping control system with revised recurrent fuzzy neural network (RRFNN) and mended particle swarm optimization (MPSO) is proposed to operate the LPMSM drive system to raise robustness of system. The RRFNN is used to estimate the value of the external lumped force uncertainty. Moreover, the error compensation control with the error compensation mechanism is proposed to compensate the minimum reconstructed error of the error estimation law. Besides, four variable learning rates in the weights of the RRFNN are regulated by virtue of MPSO with segment regulation to speed-up parameter’s convergence. Finally, comparative performances through some tentative upshots are verified that the integral backstepping control system by virtue of RRFNN with MPSO has better control performances than those of the proposed methods for the LPMSM drive system.

Design of a robust adaptive sliding mode control using recurrent fuzzy wavelet functional link neural networks for industrial robot manipulator with dead zone

This paper addresses the problem of trajectory tracking control for industrial robot manipulators (IRMs) in the presence of external disturbances and uncertain dynamics. A novel robust adaptive recurrent fuzzy wavelet functional link neural network (RFWFLNN) controller based on dead-zone compensator is proposed in order to improve the position tracking performance. To handle the unknown dynamics of the IRMs, the robust adaptive RFWFLNNs are applied to approximate the unknown dynamics. The online learning laws and estimation of the dead zone are determined by using Lyapunov stability theory and the approximation theory. In addition, the robust SMC is applied to eliminate the estimation errors and disturbances of the IRM control system. Therefore, the RFWFLNN controller for IRMs can guarantee not only the robustness and stability but also the position tracking performance. Some simulations and experiments performed on three-link IRMs are provided to prove the robustness and effectiveness of the RFWFLNNs. The superiority of the RFWFLNN controller is also demonstrated based on comparisons with fuzzy wavelet neural networks and PID controllers.

Long term electric load forecasting based on TS-type recurrent fuzzy neural network model

It is of great demand in the power industry to use hourly information to establish the appropriate predictive model and improve the forecasting accuracy in the short/long term load forecasting processes. In this paper, we propose a new load forecasting approach to provide a more accurate and defensible forecast on the long term load forecasting process. The proposed method consists of two stages: weather station selection stage and Takagi–Sugeno (TS) fuzzy model based prediction stage. In the first stage, a new weather station selection method is developed, in which the most relevant 3 weather stations are combined to generate 5 temperature series for each zone and the temperature series with the least mean absolute percentage error (MAPE) is chosen as the weather station information used to forecast. At the second stage, an improved self-organizing radial basis function recurrent neural network with TS fuzzy model (ISO-TS-RBF-RFNN) is exploited in view of the conditions such as the appearing uncertainties of holiday feature component in different years, the violation of Gaussian distribution on feature component and the temperature measurement divergence. In the proposed ISO-TS-RBF-RFNN model, a computing method on current firing strength of the fuzzy modes is developed by using a novel type of activation mechanism and robust-type fuzzy rules to improve the robustness of the predictive model. Finally, we use the same partition in the dataset from GEFCom2012 to train and demonstrate the application of the proposed model. Compared with the other five models i.e., MLR, SVR, FIR, GA-LSTM and traditional SO-TS-RBF-RFNN, the proposed model has higher forecasting accuracy than others under four evaluation criteria.

A fuzzy recurrent neural network for driver fatigue detection based on steering-wheel angle sensor data

The study of the robust fatigue feature learning method for the driver’s operational behavior is of great significance for improving the performance of the real-time detection system for driver’s fatigue state. Aiming at how to extract more abstract and deep features in the driver’s direction operation data in the robust feature learning, this article constructs a fuzzy recurrent neural network model, which includes input layer, fuzzy layer, hidden layer, and output layer. The steering-wheel direction sensing time series sends the time series to the input layer through a fixed time window. After the fuzzification process, it is sent to the hidden layer to share the weight of the hidden layer, realize the memorization of the fatigue feature, and improve the feature depth capability of the steering wheel angle time sequence. The experimental results show that the proposed model achieves an average recognition rate of 87.30% in the fatigue sample database of real vehicle conditions, which indicates that the model has strong robustness to different subjects under real driving conditions. The model proposed in this article has important theoretical and engineering significance for studying the prediction of fatigue driving under real driving conditions.

Online learning based on adaptive learning rate for a class of recurrent fuzzy neural network

This paper proposes a novel structure of a recurrent interval type-2 TSK fuzzy neural network (RIT2-TSK-FNN) controller based on a reinforcement learning scheme for improving the performance of nonlinear systems using a less number of rules. The parameters of the proposed RIT2-TSK-FNN controller are leaned online using the reinforcement actor–critic method. The controller performance is improved over the time as a result of the online learning algorithm. The controller learns from its own mistakes and faults through the reward and punishment signal from the external environment and seeks to reinforce the RIT2-TSK-FNN controller parameters to converge. In order to obtain less number of rules, the structure learning is performed and thus the RIT2-TSK-FNN rules are obtained online based on the type-2 fuzzy clustering. The online adaptation of the proposed RIT2-TSK-FNN controller parameters is developed using the Levenberg–Marquardt method with adaptive learning rates. The stability analysis is discussed using the Lyapunov theorem. The obtained results show that the proposed RIT2-TSK-FNN controller using the reinforcement actor–critic technique is more preferable than the RIT2-TSK-FNN controller without the actor–critic method under the same conditions. The proposed controller is applied to a nonlinear mathematical system and an industrial process such as a heat exchanger to clarify the robustness of the proposed structure.

Adaptive ORFWNN-Based Predictive PID Control

This paper proposes a new adaptive proportional–integral–derivative (PID) control method using predictive control and output recurrent fuzzy wavelet neural network (ORFWNN) for a group of nonlinear digital time-delay dynamic systems. The presented controller, called ORFWNN-APPID controller, is rigorously derived and proved by including an ORFWNN identifier with online parameter learning and identification, and an adaptive ORFWNN-based predictive PID controller to achieve precise setpoints tracking and disturbance rejection. The effectiveness and superiority of the constructed ORFWNN-APPID control approach are well demonstrated by performing numerical simulations on step-like disturbance rejection and precise setpoint tracking of two well-known digital nonlinear processes. The practicability of the presented method is illustrated by carrying out two experimental results on a real PET temperature control process.

An Adaptive Fuzzy Recurrent Neural Network for Solving the Nonrepetitive Motion Problem of Redundant Robot Manipulators

In order to effectively decrease the joint-angular drifts and end-effector position accumulation errors, a novel adaptive fuzzy recurrent neural network (AFRNN) is proposed and exploited to solve the nonrepetitive motion problem of redundant robot manipulators in this paper. First, a quadratic programming (QP)-based repetitive motion scheme is designed according to the kinematics constraint of redundant robot manipulators. Second, the QP-based repetitive motion scheme is converted to a matrix equation according to the Lagrangian multiplier method. Third, inspired by the neural-dynamic and fuzzy control theory, the AFRNN model is designed, which can effectively solve the matrix equation as well as the original nonrepetitive motion problem of redundant robot manipulators. Computer simulation results verify the effectiveness, high accuracy, and robustness to resist external disturbance of the proposed AFRNN scheme.

Backstepping control and revamped recurrent fuzzy neural network with mended ant colony optimization applied in SCRIM drive system

Backstepping control and revamped recurrent fuzzy neural network with mended ant colony optimization applied in SCRIM drive system

Prediction of Flyrock in Mine Blasting: A New Computational Intelligence Approach

Blasting is the predominant rock fragmentation technique in civil constructions, underground and surface mines. Flyrock is the unwanted throw of rock fragments during blasting and is the major cause of considerable damage in and around the mines. The present research aimed to propose a new intelligence-based method to predict flyrock. In this regard, the recurrent fuzzy neural network (RFNN) combined with the genetic algorithm (GA) is proposed. For checking the suitability of the RFNN-GA model, artificial neural network (ANN), hybrid ANN and GA and a nonlinear regression model were also employed. To achieve the aims of the research, data for 70 blasting sites including four input parameters (spacing, burden, stemming and maximum charge per delay) and one output parameter (flyrock) were gathered from two quarry mines at the Shur River dam, Iran. The performance of the proposed prediction methods was then assessed with statistical evaluation criteria, i.e., R-square and root mean square error. The results indicate the proposed RFNN-GA model was more superior for prediction of flyrock than the GA-ANN, ANN and nonlinear regression models. According to a sensitivity analysis, the maximum charge per delay was the most influential parameter in flyrock prediction in this case.

Motion Planning of Autonomous Mobile Robot Using Recurrent Fuzzy Neural Network Trained by Extended Kalman Filter.

This paper proposes a novel motion planning method for an autonomous ground mobile robot to address dynamic surroundings, nonlinear program, and robust optimization problems. A planner based on the recurrent fuzzy neural network (RFNN) is designed to program trajectory and motion of mobile robots to reach target. And, obstacle avoidance is achieved. In RFNN, inference capability of fuzzy logic and learning capability of neural network are combined to improve nonlinear programming performance. A recurrent frame with self-feedback loops in RFNN enhances stability and robustness of the structure. The extended Kalman filter (EKF) is designed to train weights of RFNN considering the kinematic constraint of autonomous mobile robots as well as target and obstacle constraints. EKF's characteristics of fast convergence and little limit in training data make it suitable to train the weights in real time. Convergence of the training process is also analyzed in this paper. Optimization technique and update strategy are designed to improve the robust optimization of a system in dynamic surroundings. Simulation experiment and hardware experiment are implemented to prove the effectiveness of the proposed method. Hardware experiment is carried out on a tracked mobile robot. An omnidirectional vision is used to locate the robot in the surroundings. Forecast improvement of the proposed method is then discussed at the end.

A novel intelligent approach to simulate the blast-induced flyrock based on RFNN combined with PSO

This research focuses to propose a new hybrid approach which combined the recurrent fuzzy neural network (RFNN) with particle swarm optimization (PSO) algorithm to simulate the flyrock distance induced by mine blasting. Here, this combination is abbreviated using RFNN–PSO. To evaluate the acceptability of RFNN–PSO model, adaptive neuro-fuzzy inference system (ANFIS) and non-linear regression models were also used. To achieve the objective of this research, 72 sets of data were collected from Shur river dam region, in Iran. Maximum charge per delay, stemming, burden, and spacing were considered as input parameters in the models. Then, the performance of the RFNN–PSO model was evaluated against ANFIS and non-linear regression models. Correlation coefficient (R2), Nash and Sutcliffe (NS), mean absolute bias error (MABE), and root-mean-squared error (RMSE) were used as comparing statistical indicators for the assessment of the developed approach’s performance. Results show a satisfactory achievement between the actual and predicted flyrcok values by RFNN–PSO with R2, NS, MABE, and RMSE being 0.933, 0.921, 13.86, and 15.79, respectively.

Novel Fuzzy Neural Nonsingular Terminal Sliding Mode Control of MEMS Gyroscope

This paper attempts to improve the robustness and rapidity of a microgyroscope sensor by presenting a double‐loop recurrent fuzzy neural network based on a nonsingular terminal sliding mode controller. Compared with the traditional control method, the proposed strategy can obtain faster dynamic response speed and lower steady‐state error with high robustness in the presence of system uncertainties and external disturbances. A nonlinear terminal sliding mode controller is designed to guarantee finite‐time high‐precision convergence of the sliding surface and meanwhile to eliminate the effect of singularity. Moreover, an exponential approach law is used to accelerate the convergence rate of the system to the sliding surface. For suppressing the chattering, the symbolic function in the ideal sliding mode is replaced by the saturation function. To suppress the effect of model uncertainties and external disturbances, a double‐loop recurrent fuzzy neural network is introduced to approximate and compensate system nonlinearities for the gyroscope sensor. At the same time, the double‐loop recurrent fuzzy neural network can effectively accelerate the speed of parameter learning by introducing the adaptive mechanism. Simulation results indicate that the control system with the proposed controller is easily implemented, and it has higher tracking precision and considerable robustness to model uncertainties compared with the existing controllers.

Using self-constructing recurrent fuzzy neural networks for identification of nonlinear dynamic systems

In this paper, the self-constructing recurrent fuzzy neural network (SCRFNN) is applied for nonlinear dynamical system identification (NDSI). The SCRFNN is a novel fuzzy neural network (FNN) by adding a recurrent path in each node of the hidden layer of self-constructing FNN, which contains two learning phases. Specifically, the structure learning is based on partition of the input space and the parameter learning is based on the supervised gradient descent method using a delta adaptation law. The SCRFNN can decrease the minimum firing strength in each learning cycle and the number of hidden neurons which is an FNN with high accuracy and compact structure compared with several other neural networks. The performance of SCRFNN in NDSI is further verified in simulation.

Modified PSO Algorithm on Recurrent Fuzzy Neural Network for System Identification

Modified PSO Algorithm on Recurrent Fuzzy Neural Network for System Identification

Using self-constructing recurrent fuzzy neural networks for identification of nonlinear dynamic systems

Using self-constructing recurrent fuzzy neural networks for identification of nonlinear dynamic systems

Online identification of a rotary wing Unmanned Aerial Vehicle from data streams

Until now the majority of the neuro and fuzzy modeling and control approaches for rotary wing Unmanned Aerial Vehicles (UAVs), such as the quadrotor, have been based on batch learning techniques, therefore static in structure, and cannot adapt to rapidly changing environments. Implication of Evolving Intelligent System (EIS) based model-free data-driven techniques in fuzzy system are good alternatives, since they are able to evolve both their structure and parameters to cope with sudden changes in behavior, and performs perfectly in a single pass learning mode which is suitable for online real-time deployment. The Metacognitive Scaffolding Learning Machine (McSLM) is seen as a generalized version of EIS since the metacognitive concept enables the what-to-learn, how-to-learn, and when-to-learn scheme, and the scaffolding theory realizes a plug-and-play property which strengthens the online working principle of EISs. This paper proposes a novel online identification scheme, applied to a quadrotor using real-time experimental flight data streams based on McSLM, namely Metacognitive Scaffolding Interval Type 2 Recurrent Fuzzy Neural Network (McSIT2RFNN). Our proposed approach demonstrated significant improvements in both accuracy and complexity against some renowned existing variants of the McSLMs and EISs.

System Identification and Prediction of Dynamic System and Microwave Thermal Process Using a Recurrent Fuzzy Quantum Neural Network

Microwave heating is a time-varying, non-linear process. Mechanism modeling of the microwave thermal process is extremely difficult because of the complex microwave heating environment. This paper presents a recurrent fuzzy quantum neural network with full feedbacks (RFQNN) for prediction and identification of dynamic systems and the actual microwave heating process. In the RFQNN, a quantum neural network is introduced to the consequent part of the fuzzy rules to improve the mapping ability and the identification precision. All of the rules are generated and learned online through a simultaneous structure and parameter learning. During the structure learning, an online clustering algorithm combined with Mahalanobis distance elimination algorithm perform effectively in generating or removing fuzzy rules. And then a gradient descent algorithm is introduced to update the parameters during the parameter learning process. And finally, we test the RFQNN by dynamic plants and the microwave thermal process. The results show that it performs well in dynamic system processing compared with other recurrent fuzzy neural networks.

Mean-Square Exponential Input-to-State Stability of Stochastic Fuzzy Recurrent Neural Networks with Multiproportional Delays and Distributed Delays

We are interested in a class of stochastic fuzzy recurrent neural networks with multiproportional delays and distributed delays. By constructing suitable Lyapunov-Krasovskii functionals and applying stochastic analysis theory, Ito^’s formula and Dynkin’s formula, we derive novel sufficient conditions for mean-square exponential input-to-state stability of the suggested system. Some remarks and discussions are given to show that our results extend and improve some previous results in the literature. Finally, two examples and their simulations are provided to illustrate the effectiveness of the theoretical results.