Diagonal Recurrent Neural Network Research Articles

Development of sliding mode control based on diagonal recurrent neural network for coupled tank system

This study presents the development of sliding mode control (SMC) using the diagonal recurrent neural network (DRNN) for nonlinear systems. Firstly, the SMC for linear systems is developed for nonlinear coupled tank system. Second, the DRNN is used to design the equivalent part of the SMC law, which is performed to approximate the dynamics of a controlled process. Third, the sliding surface for the switching control is developed using the DRNN. The DRNN parameters are tuned using Lyapunov function to achieve the controlled process stability. For the developed scheme, discontinuous signum function is used to compensate the chattering phenomenon. The developed scheme is applied for controlling the uncertain nonlinear coupled tank system. The simulation results indicate that the developed scheme can respond to the effects of system uncertainties compared to other existing schemes.

Parameter Identification of Nonlinear Flux-Linkage Model for Switched Reluctance Motor Based on Chaotic Diagonal Recurrent Neural Network

Parameter Identification of Nonlinear Flux-Linkage Model for Switched Reluctance Motor Based on Chaotic Diagonal Recurrent Neural Network

Proxy-Based Sliding Mode Force Control for Compliant Grinding via Diagonal Recurrent Neural Network and Prandtl-Ishlinskii Hysteresis Compensation Model

Traditional industrial robots often face challenges in achieving a perfectly polished surface on a workpiece because of their high mechanical rigidity. The active compliance force control device installed at the robotic arm’s end enables high-precision contact force control between the grinding tool and the workpiece. However, the complex hysteresis nonlinearity between cylinder air pressure and output force, as well as various random disturbances during the grinding process, can affect the accuracy of the contact force and potentially impact the grinding effect of the workpiece, even causing irreversible damage to the surface of the workpiece. Given the complex random variation of cylinder output force in the actual grinding process, a rate-dependent hysteresis model based on diagonal recurrent neural network and Pradtl–Ishlinskii models named dRNN-PI is designed to compensate for the complex nonlinear hysteresis of the cylinder and calculate the desired air pressure to maintain a steady contact force on the workpiece. The proxy-based sliding mode control (PSMC) is utilized to quickly track the desired air pressure without overshooting. This paper also proves the controller’s stability using the Lyapunov-based methods. Finally, the accuracy of the proposed hysteresis compensation model and the effectiveness and robustness of the PSMC are verified by experiment results.

Operation Control Method for High-Speed Maglev Based on Fractional-Order Sliding Mode Adaptive and Diagonal Recurrent Neural Network

The speed profile tracking calculation of high-speed maglev trains is mainly affected by running resistance. In order to reduce the adverse effects and improve tracking accuracy, this paper presents a maglev train operation control method based on a fractional-order sliding mode adaptive and diagonal recurrent neural network (FSMA-DRNN). First, the kinematic resistance equation is established due to the three types of resistance that occur during the actual operation of a train: air resistance, guide eddy current resistance, and suspension frame generator coil resistance. Then, the FSMA-DRNN control law and parameter update law are designed, and a FSMA-DRNN operation controller is composed of three parts: speed feed forward, fractional-order sliding mode adaptive equivalent control, and diagonal recurrent neural network resistance compensation. Furthermore, by using the designed operation controller, it is proven effective by the Lyapunov theory for the stability of the closed-loop control system. Apart from the proposed theoretical analysis, the proposed approaches are verified by experiments on the high-speed maglev hardware-in-the-loop simulation platform Rt-Lab, in line with the 29.86 km test line and a five-car train from the Shanghai maglev, showing the effectiveness and superiority for operation optimization.

Validation of target tracking performance through signal feature extraction method based on 1D convolutional neural network

In order to perform active vibration and noise control, a variety of adaptive algorithms have been investigated and developed. Especially, neural network-based signal tracking algorithms, such as the radial basis function neural network (RBFNN), are being widely utilized. For signal tracking problems, a reference signal has a significantly important role, while it is difficult to determine the reference signal when it has relatively complex spectrum. Thus, this study focuses on the signal feature extraction for manipulating appropriate reference signals easily and validation of signal tracking performance with it. In order to carried out the signal feature extraction, 1D convolutional neural network (1D CNN) is implemented and trained based on the CWRU bearing dataset. In addition, multi normalized least mean square (NLMS), neural network-based signal tracking algorithm, and diagonal recurrent neural network (DRNN) is employed to confirm the signal tracking performance. The proposed method shows that the 1D CNN-based signal feature extraction method could find the signal feature properly and the reference signal obtained from this methodology is employed to a target tracking problem, which shows a great performance.

Linear Active Disturbance Rejection Control-Based Diagonal Recurrent Neural Network for Radar Position Servo Systems with Dead Zone and Friction

This paper proposes a control scheme for the radar position servo system facing dead zone and friction nonlinearities. The controller consists of the linear active disturbance rejection controller (LADRC) and diagonal recurrent neural network (DRNN). The LADRC is designed to estimate in real time and compensate for the disturbance with vast matched and mismatched uncertainties, including the internal dead zone and friction nonlinearities and external noise disturbance. The DRNN is introduced to optimize the parameters in the linear state error feedback (LSEF) of the LADRC in real time and estimate the model information, namely Jacobian information, of the plant on-line. In addition, considering the Cauchy distribution, an adaptive tracking differentiator (ATD) is designed in order to manage the contradiction between filtering performance and tracking speed, which is introduced to the LADRC. Another novel idea is that the back propagation neuron network (BPNN) is also introduced to tune the parameters of the LADRC, just as in the DRNN, and the comparison results show that the DRNN is more suitable for high precision control due to its feedback structure compared with the static BPNN. Moreover, the regular controller performances and robust performance of the proposed control approach are verified based on the radar position servo system by MATLAB simulations.

Diagonal Recurrent Neural Network-Based Hysteresis Modeling.

The Preisach model and the neural networks are two of the most popular strategies to model hysteresis. In this article, we first mathematically prove that the rate-independent Preisach model is actually a diagonal recurrent neural network (dRNN) with the binary step activation function. For the first time, the hysteresis nature and conditions of the classical dRNN with the tanh activation function are mathematically discovered and investigated, instead of using the common black-box approach and its variants. It is shown that the dRNN neuron is a versatile rate-dependent hysteresis system under specific conditions. The dRNN composed of those neurons can be used for modeling the rate-dependent hysteresis and it can approximate the Preisach model with arbitrary precision with specific parameters for rate-independent hysteresis modeling. Experiments show that the classical dRNN models both kinds of hysteresis more accurately and efficiently than the Preisach model.

Recurrent-Neural-Network-Based Rate-Dependent Hysteresis Modeling and Feedforward Torque Control of the Magnetorheological Clutch

The complicated rate-dependent hysteresis due to the input current and rotation speed of the magnetorheological clutch (MRC) greatly hinders its applications in robotics. In this article, it is first proved that the Preisach model is a special diagonal recurrent neural network (dRNN) with a binary step activation function. Motivated by this, the condition and the internal knowledge of the classical dRNN in modeling the rate-dependent hysteresis are investigated and validated. To guarantee that such a condition holds in dRNN training, a new loss function is proposed. The coupled rate-dependent hysteresis of an MRC prototype is, then, precisely modeled by the dRNN with the current input and speed input. By employing the trained dRNN model, the MRC torque is controlled using the feedforward scheme with no external torque sensors. Taking the advantage of being free-of-parameter tuning for the classical inverse multiplicative structure control (IMSC), a dRNN-based IMSC is proposed by finding the best feasible neuron and solving a nonlinear function. Various experiments validate much better control accuracy and higher bandwidth of the IMSC over the model-based proportional–integral–derivative control.

Hybrid deep learning diagonal recurrent neural network controller for nonlinear systems

In the present paper, a hybrid deep learning diagonal recurrent neural network controller (HDL-DRNNC) is proposed for nonlinear systems. The proposed HDL-DRNNC structure consists of a diagonal recurrent neural network (DRNN), whose initial values can be obtained through deep learning (DL). The DL algorithm, which is used in this study, is a hybrid algorithm that is based on a self-organizing map of the Kohonen procedure and restricted Boltzmann machine. The updating weights of the DRNN of the proposed algorithm are developed using the Lyapunov stability criterion. In this concern, simulation tasks such as disturbance signals and parameter variations are performed on mathematical and physical systems to improve the performance and the robustness of the proposed controller. It is clear from the results that the performance of the proposed controller is better than other existent controllers.

A Lyapunov-stability-based context-layered recurrent pi-sigma neural network for the identification of nonlinear systems

A novel higher-order context-layered recurrent pi-sigma neural network (CLRPSNN) is presented for the identification of nonlinear dynamical systems. The proposed model is the modified form of the classical pi-sigma neural network (PSNN) and contains an additional layer (known as the context layer) of the context nodes. Pi-sigma networks involve a product operator/unit in their output layer which indirectly incorporates in them the capability of higher-order networks and also reduces their network complexity. For tuning the weights of the proposed CLRPSNN model, a learning procedure is developed by combining the Back-Propagation (BP) and Lyapunov-stability method. The performance of the proposed model is compared with other models such as PSNN, Feed-forward neural network (FFNN) (containing single hidden layer), and various popular recurrent neural network (RNN) like Elman recurrent neural network (ERNN), Jordan recurrent neural network (JRNN), Diagonal recurrent neural network (DRNN), and a deep neural network (DNN). The simulation study showed that the proposed model has given better results as compared to the other models.

Nonlinear model predictive control based on hyper chaotic diagonal recurrent neural network

Nonlinear model predictive control based on hyper chaotic diagonal recurrent neural network

HCDRNN‐NMPC: A New Approach to Design Nonlinear Model Predictive Control (NMPC) Based on the Hyper Chaotic Diagonal Recurrent Neural Network (HCDRNN)

In industrial applications, Stewart platform control is especially important. Because of the Stewart platform’s inherent delays and high nonlinear behavior, a novel nonlinear model predictive controller (NMPC) and new chaotic neural network model (CNNM) are proposed. Here, a novel NMPC based on hyper chaotic diagonal recurrent neural networks (HCDRNN‐NMPC) is proposed, in which, the HCDRNN estimates the future system’s outputs. To improve the convergence of the parameters of the HCDRNN to better the system’s modeling, the extent of chaos is adjusted using a logistic map in the hidden layer. The proposed scheme uses an improved gradient method to solve the optimization problem in NMPC. The proposed control is used to control six degrees of freedom Stewart parallel robot with hard‐nonlinearity, input constraints, and in the presence of uncertainties including external disturbance. High prediction performance, parameters convergence, and local minima avoidance of the neural network are guaranteed. Stability and high tracking performance are the most significant advantages of the proposed scheme.

DRNN Robust DTC for Induction Motor Drive Systems Using FSTPI

In this paper, a self-tuning PI speed controller based on diagonal recurrent neural network is (DRNN) investigated and simulated to increase the robustness of the direct torque control (DTC) scheme for three-phase low-power IM drive system using a Four Switch Three-Phase Inverter (FSTPI). The drive is subjected to different system inputs and disturbances, step changes in speed under different load conditions, abrupt loading at high speed and speed reversal. Furthermore, the robustness of the controller is evaluated by varying motor parameter, stator resistance and moment of inertia. A comparison of classical and self-tuning PI speed controllers was presented to determine the effectiveness of the proposed controller. It is concluded based on simulation results using Matlab/Simulink. that the self-tuning PI speed controller provides the best performance by reacting rapidly and adaptively.

Full-state neural network observer-based hybrid quantum diagonal recurrent neural network adaptive tracking control

This study introduces a neural network (NN) adaptive tracking controller-based reinforcement learning (RL) scheme for unknown nonlinear systems. First, an observer using feed-forward NN (FFNN) is performed to estimate the controlled system states. Second, an adaptive control based on actor-critic RL is developed, in which a quantum diagonal recurrent neural network (QDRNN) is proposed to represent the critic and actor parts. The critic QDRNN is applied to perform the “strategic” utility function, and it is minimized by the actor QDRNN. The proposed adaptive tracking NN control guarantees the faster convergence due to the developed updated algorithm for the controller parameters, which is derived using the Lyapunov function. Simulation and practical results indicate the robustness of the proposed observer-based adaptive control relative to other existing controllers.

Recurrent neural network-based prediction of compressive and flexural strength of steel slag mixed concrete

In this paper, an effort has been made to develop a recurrent type of neural network known as diagonal recurrent neural network (DRNN) to predict the compressive and flexural strengths of AOD steel slag-mixed concrete for pavements. The data used for modeling were attained from the laboratory experiments. The compressive and flexural strengths were experimentally analyzed for specimens containing 0%, 10%, 15%, 20%, and 25% of AOD steel slag as a partial replacement of cement at curing ages of 3, 7, 28, 90, 180, and 365 days. The developed model was trained using the backpropagation (BP) algorithm. The performance of the proposed model during the training and validation has been compared with the well-known prediction models such as multi-layer perceptron (MLP) and the radial basis function network (RBFN). The DRNN-based prediction model has given much better prediction results when compared to the other two models since the former provided comparatively smaller values of performance indicators such as average mean square error (AMSE) and mean average error (MAE). The reason for DRNN performing better than the other two models is that it contains feedback connections/weights which induce memory property in its structure. This helps DRNN to better model the complex mappings. Such feedback loops are not available in MLP and RBFN. The study conducted in this research concludes that the DRNN-based prediction model should be preferred over the MLP and RBFN models for predicting the compressive and flexural strengths of AOD steel slag added to concrete for pavements.

Q-learning optimized diagonal recurrent neural network control strategy for brushless direct current motors

In order to improve the working stability of brushless direct current motors (BLDCM), a diagonal recursive neural network (DRNN) control strategy based on Q-learning algorithm is proposed in this paper which is called as Q-DRNN. In Q-DRNN, DRNN iterates over the output variables through a unique recursive loop in the hidden layer, and its key weight is optimized to speed up the iteration. Moreover, an improved Q-learning algorithm is introduced to modify the weight momentum factor of DRNN, which makes DRNN have the ability of learning and online correction so as to make the BLDCM achieve better control effect. In MATLAB/Simulink environment, Q-DRNN is tested and compared with other popular control methods in terms of speed and torque response under different operating conditions, and the results show that Q-DRNN has better adaptive and anti-interference ability as well as stronger robustness.

Diagonal recurrent neural network observer-based adaptive control for unknown nonlinear systems

This paper proposes an observer-based adaptive control for unknown nonlinear systems using an adaptive dynamic programming (ADP) algorithm. First, a diagonal recurrent neural network (DRNN) observer is proposed to estimate the unknown dynamics of the nonlinear system states. The proposed neural network offers a simpler structure with deeper memory and guarantees the faster convergence. Second, a neural controller is constructed via ADP method using the observed states to get the optimal control. The optimal control law is determined based on the new structure of the critic network, which is performed using the DRNN. The learning algorithm for the proposed DRNN observer-based adaptive control is developed based on the Lyapunov stability theory. Simulation results and hardware-in-the-loop results indicate the robustness of the proposed ADP to respond the system uncertainties and external disturbances compared with other existing schemes.

NARMAX Model-Based Hysteresis Modeling of Magnetic Shape Memory Alloy Actuators

Magnetic shape memory alloy (MSMA) based actuators are extensively applied in the fields of precision manufacturing and micro/nano technology. Nevertheless, the inherent hysteresis in the MSMA-based actuator severely hinders its further application. In this letter, the characteristics of hysteresis behavior under different input signals are investigated. Then, a nonlinear auto-regressive moving average with exogenous inputs (NARMAX) model based on a diagonal recurrent neural network (DRNN) is used to construct the rate-dependent hysteresis model. To improve the capability of characterizing the multi-valued mapping of the hysteresis loop, the play operator is adopted as the exogenous variable function of the NARMAX model. To verify the effectiveness of the proposed model, a series of comparisons are implemented. The experimental results show that the proposed NARMAX model based on the DRNN exhibits excellent modeling performance.

Predictive control of slurry pressure balance in shield tunneling using diagonal recurrent neural network and evolved particle swarm optimization

Abstract Establishing the balance between slurry supporting pressure and expected water-earth pressure is an important criterion to ensure excavating face stability in shield tunneling. To overcome the inaccuracy and hysteresis of manual operations, this paper presents a model predictive control (MPC) system for the slurry pressure balance during construction through effectively regulating the slurry circulation and air pressure holding systems according to geological conditions. The MPC structure consists of a diagonal recurrent neural network (DRNN) that approximates the complex relationship between slurry pressure and tunneling parameters, an optimizer which produces the optimal air pressure and slurry level based on the multi-step ahead predictions, and an evolved particle swarm optimization (EPSO) algorithm. The proposed EPSO can update the structure and weights of DRNN concurrently to better cater to the changeable stratum. The optimizer can excellently compensate the time delays in slurry pressure regulation by incorporating the logical control sequence of actuator systems into the EPSO procedure. The simulation results demonstrated that the presented approach can accurately track the desired water-earth pressure and significantly enhance the robustness of slurry supporting system in tunneling, and the novel EPSO also performed higher convergence speed and precision than the classic algorithms used for comparison.

Trajectory planning and diagonal recurrent neural network vibration control of a flexible manipulator using structural light sensor

Abstract A new vibration measurement method and active control algorithm for a flexible manipulator are investigated. A non-contact vibration measurement method based on structural light sensor is proposed. During and after the point-to-point motion, the locally controlled autoregressive moving average (CARMA) model of the flexible manipulator is determined through experiments. Based on the model, the particle swarm optimization (PSO) algorithm is adopted to obtain the optimal vibration suppression trajectory. Moreover, considering the time-varying and nonlinear characteristics of the flexible manipulator, a diagonal recurrent neural network (DRNN) control algorithm is designed to suppress residual vibration, which consists of an on-line identifier and a vibration control signal generator. The experiment setup is constructed. Compared with trapezoidal trajectory and classic smooth trajectories, planning an optimal trajectory can cause less vibration under the same motion requirements. Experimental results demonstrated that the DRNN controller is superior to the classical PD controller on vibration suppression, especially for the small amplitude residual vibration. Furthermore, the hybrid control strategy of optimal trajectory planning and DRNN control has the advantages of ensuring smoother motion and faster residual vibration suppression speed.