Neural Network-based Sliding Mode Control Research Articles

Adaptive neural network-based sliding mode control for trajectory tracking control of cable-driven continuum robots with uncertainties

Adaptive neural network-based sliding mode control for trajectory tracking control of cable-driven continuum robots with uncertainties

Intelligent and robust control of space manipulator for sustainable removal of space debris

This study focuses on enhancing the control precision and efficiency of a two-degree-of-freedom (2-DOF) space manipulator used for active space debris removal. The unpredictable space environment introduces large uncertainties, which introduces unique challenges beyond the capabilities of a standalone computed torque controller and degrades control performance. To address this problem, a robust controller is developed, integrating traditional techniques such as sliding mode and computed torque control with a Neural Network framework. This synergy leverages both methods' strengths—conventional controls' accuracy and Neural Network's adaptability. The integration of Neural Network-based sliding mode control complements the robustness of computed torque control by actively mitigating uncertainties and disturbances inherent in the space environment. The 2-DOF manipulator's state variables model the system dynamics, necessitating accurate relative motion estimation between the manipulator and debris. The global asymptotic stability of the developed algorithm is demonstrated through the Lyapunov theorem, guaranteeing error convergence to zero. The convergence, stability, precision, tracking errors, and responsiveness of the controller have been analysed and validated by the MATLAB Simulink simulations. The novel approach's performance effectiveness is substantiated by numerical simulations and a comparative analysis with conventional computed torque control. Outcomes highlight the superior precision and efficiency in manipulator tracking the trajectory, validating the integrated controller's potential for successful active space debris removal.

Adaptive neural network-based sliding mode control of rotary inverted pendulum system

This paper introduces a new adaptive neural network-based sliding mode control of rotary inverted pendulum system (RIPS) subject to model uncertainties and disturbances. First, the sliding mode control is designed using the double power rate reaching law. Second, an adaptive neural network (ANN) is exploited to deal with model uncertainties and disturbances. The weight matrices of the neural network are adaptively tuned. In the proposed control method the exact dynamic model of the RIPS is not required due to the exploitation of the universal property of the neural network. Moreover, the implementation of a double power-reaching law can greatly suppress the undesirable chattering. The convergence and stability analysis is validated using Lyapunov stability theory. Finally, the robustness and excellence of the proposed control strategy are verified by comparing the system performance with the existing SMC method and improved super-twisting SMC method.

An Application of Neural Network-based Sliding Mode Control for Multilevel Inverters

Multi-level 3-phase inverters using cascaded H-bridges are becoming prominent in the electric drive and renewable energy sectors due to their high capacity and ability to withstand high voltage shocks. Therefore, the modulation and control techniques used in these multilevel inverters have a crucial influence on the quality of the output voltage they produce. The significantly high common-mode voltage amplitude they generate is one of their disadvantages, causing leakage currents and harmonics. This article proposes a new technique using sliding mode control combined with neural networks to manage a three-phase multi-level inverter. The research objective of this innovative technique is to eliminate the need for current controllers and conventional modulation that relies on carrier signals, reducing hardware calculations and enhancing dynamic response. In addition, it demonstrates the ability to minimize harmonics, common mode voltage, and the number of switching counts, thereby limiting the inverter switching losses and increasing device performance. Simulation results performed on a 5-level 3-phase inverter using cascaded H-bridges have confirmed the effectiveness of the proposed method.

Artificial Neural Network-Based Experimental Investigations for Sliding Mode Control of an Induction Motor in Power Steering Applications

Sliding mode control (SMC) of induction motor is a new concept in the current scenario, as it seeks to improve torque control accuracy and power steering efficiency through the use of pulse width modulation schemes. Furthermore, artificial neural network-based sliding mode control is applied to a squirrel cage induction motor, which is used in the steering control of automobiles. The artificial neural network (ANN)-based SMC is more popular due to its robustness and good stability in external parameter variation. Additionally, an SMC and an ANN-based SMC are employed to compute the torque and flux, improving performance for power steering applications. The performance of the designed model is validated through MATLAB/Simulink and experimental models with different controllers under various operating conditions. The controller has been embedded into a TMS320F28335 controller, and performances have been evaluated. The performance analyses of induction motor using different controllers are performed. The transient performances of induction motor such as delay time, rise time, settling time, and steady-state error are investigated. The proposed work is analysed by using a mathematical model and implemented in a test-bench model for validation.

Sliding mode control based on RBF neural network for a class of underactuated systems with input quantization and event-triggering

In network control, signal transmission between each system component is carried out through the communication network. Since the bandwidth of the network is limited, quantization is a vital and fundamental technology used to convert the continuous signal to an approximate signal with a finite number of discrete value levels. Furthermore, event-triggering is an effective method to reduce signal transmission frequency in network control. Input signal quantization and event-triggering are considered simultaneously in this study for a class of underactuated systems. First, the logarithmic quantizer is used to quantize the input signal, and then the quantized input signal is further processed by the event-triggered mechanism based on the fixed threshold strategy. Adopting the proposed adaptive control scheme with the aid of radial basis function (RBF) neural network-based sliding mode control, the states of the closed-loop system are guaranteed to be bounded, and the control goals can be achieved. Finally, the control effect is shown through the numerical simulations.

Data‐based learning control for optimization of nonlinear systems

Data‐based learning control for optimization of nonlinear systems

Switched Controller Design for Robotic Manipulator via Neural Network-Based Sliding Mode Approach

In this brief, a novel adaptive switched controller of uncertain industry robotic manipulator (IRM) systems with switching loads is designed for trajectory tracking by utilizing switching neural network-based sliding mode control (SNNSMC) scheme. However, the error accumulation and system instability of the IRM may be caused by switching loads, while the existing single adaptive controller may not effectively deal with this issue. To do so, the IRM with switching loads is modeled as a switched system with multi-modal. Besides, the individual sub-controller with adaptive gain algorithm is devised to replace the single adaptive controller with fixed-gain for multi-modality system. Then, the radial basis function NN (RBFNN) is utilized for approaching to the plant, which avoids the limitation of the accurate model for the IRM system with switching loads. The SMC gain is designed as an adaptive adjustment value, which can be adjusted in real time to enhance the robustness of the system in spite of unknown disturbances and uncertainties. Subsequently, based on average dwell time (ADT) principle, the trajectory tracking error being close to zero is verified by the multi-Lyapunov function method. Finally, simulation results show that the provided method can both track the preset trajectory accurately, and attenuate chattering and high-speed switching of the plant effectively.

Adaptive neural network-based sliding mode control for a hydraulic rotary drive joint

Hydraulic system has specific nonlinear and unknown modeling characteristics, realizing precise tracking control is a very challenging task. In this study, a sliding mode controller based on an adaptive higher-order neural network is proposed for realizing precise position-tracking control of a hydraulic rotary drive joint. First, the structural design and working principle of the target joint are introduced, and the mathematical model of the corresponding valve-controlled hydraulic position servo system is developed. Then, the adaptive neural network algorithm and sliding mode control are effectively combined. Based on the measurement information obtained from the control process, the feedback error is used to adaptively approximate the control parameters and realize online adjustment of the controller output. Finally, the parameter identification and position tracking of the system are validated, and the results indicate that the proposed strategy exhibits a 20% improvement in position-tracking control compared to the conventional sliding mode control methods.

Adaptive Neuro Sliding Mode Control of Superconducting Magnetic Energy Storage System

ABSTRACT In this paper, an adaptive RBF neural network-based sliding mode controller is developed for superconducting magnetic energy storage (SMES) installed in a wind–diesel power system. Due to sudden load and wind power variations in a wind–diesel power system, power imbalances may force the system frequency to deviate from its nominal value and drive the system to operate in an unstable mode. Therefore, in a wind–diesel power system using a converter interface, a fast-acting and high power density SMES device is interconnected to carry the required power exchange whenever power imbalances occur. Based on switching manifold design, a sliding mode controller is developed to control the charging and discharging operations of SMES coil as per the power requirements, and the same is achieved by controlling the converter. A neural network using a radial basis function (RBF) is developed to estimate the unknown function of the system. Lyapunov stability analysis is conducted to guarantee the asymptotic stability of the system. MATLAB simulations are carried out and are presented to show the improved performance with the system exposed to disturbances in load and wind power.

Neural network-based sliding mode control for satellite attitude tracking

This paper studies the intelligent sliding mode control for satellite dynamics based on neural learning and disturbance observer. We consider the dynamics uncertainty due to uncertain inertia matrix and the disturbance torque caused by orbit transfer engine. A neural network is used to approximate the unknown dynamics, while an adaptive observer is utilized to handle the unknown disturbance. The prediction error is constructed by introducing the estimation model to understand the uncertainties better. Based on the finite-time design, an efficient controller is designed to achieve the attitude tracking. The system stability is guaranteed under the proposed control design while the sliding mode converges in finite time. Better learning and control performances are presented by conducting the simulation studies.

Direct adaptive neural network-based sliding mode control of a high-speed, ultratall building elevator using genetic algorithm

A direct adaptive sliding mode controller (SMC) based on radial basis function neural network (RBFNN) approximation is proposed for a high-speed, ultratall building elevator system using genetic algorithm (GA) to optimise the control parameters. The nonlinear dynamic model of the elevator system is described, with the RBFNN used to approximate the elevator system functions and external disturbance uncertainties. The RBFNN parameters are optimised using GA. The RBFNN-SMC was compared with a traditional sliding mode controller, nonlinear pseudo-derivative feedback (NPDF) controller and a nonlinear proportional-integral-derivative controller. The Lyapunov stability theorem is applied to develop the adaptive law, thereby guaranteeing the system stability. Performance of the proposed RBFNN-SMC has been evaluated using numerical simulations. The RBFNN-SMC achieved effective control of the elevator system. Although the RBFNN-SMC system achieved comparable pre-re-levelling control to its competitors, problematic chattering was observed due to sensor noise, suggesting that the system must be coupled with a noise-attenuating filter to avoid actuator damage. Following arrival of the cabins, an adaptive re-levelling operation was applied to reduce the distance between the cabins and the arrival floor. Although both SMC variants accomplished successful re-levelling, the NPDF-based controller achieved the best performance—adjusting the final cabin position to within 1 mm of the target floor in both considered displacement overshoot cases.

Iterative learning based neural network sliding mode control for repetitive tasks: With application to a PMLSM with uncertainties and external disturbances

The uncertainties, force ripple, friction and other external disturbances in permanent magnet linear synchronous motors (PMLSMs) pose great challenges to the design of advanced controllers with high speed and high precision. Existing neural network based sliding mode control (NNSMC) methods typically adopt neural network as an online compensator for uncertainties and disturbances, meanwhile stability and robustness are assured owing to sliding mode control. However, PMLSMs usually need to perform repetitive motion tasks within a finite time interval in practical applications, where existing NNSMC methods ignore the previous repetitions’ information, and consequently limit the motion performance improvement for trajectory tracking tasks. This paper proposes an iterative learning based neural network sliding mode control (ILNNSMC) scheme, in which the concept of iterative learning is incorporated into NNSMC methods for the first time. In ILNNSMC, the uncertainties and disturbances can be compensated by NN (updating in the time-domain) and ILC (learning in the iteration-domain) in a complementary way, and the sliding variable is driven to converge to zero for arbitrary time within the predesigned time interval with successive iterations. The motion performance for repetitive tasks can be improved substantively. Moreover, a non-linear dynamic learning gain (DLG) strategy is proposed and adopted in iterative learning law, which guarantees both convergence speed and steady-state accuracy in the learning process. Stability of the scheme is discussed in both time and iteration domains, and the control performance is verified by extensive experiments on a PMLSM.

RBF Neural Network-Based Sliding Mode Control for Modular Multilevel Converter with Uncertainty Mathematical Model

For medium and high-powered applications, modular multilevel converters have become the most promising converter application. In this paper, a sliding mode controller based on an RBF neural network is proposed for a modular multilevel converter. The RBF neural network is designed to approximate the uncertainty mathematical model of a modular multilevel converter. The main innovation of the proposed method is that it does not require any model parameters and control parameters during the whole control process. This means that parameter changes caused by the external environment will not influence the controller performances. Finally, by comparing with a conventional PI controller, the simulation proves the feasibility and effectiveness of the proposed control method. In addition, the experimental results show that the grid-side current can become stable immediately while the active power is stabilized after 20 ms when the set value is changed.

Adaptive Sliding Mode Control Method for Onboard Supercapacitors System

Urban rail trains have undergone rapid development in recent years due to their punctuality, high capacity and energy efficiency. Urban trains require frequent start/stop operations and are, therefore, prone to high energy losses. As trains have high inertia, the energy that can be recovered from braking comes in short bursts of high power. To effectively recover such braking energy, an onboard supercapacitor system based on a radial basis function neural network-based sliding mode control system is proposed, which provides robust adaptive performance. The supercapacitor energy storage system is connected to a bidirectional DC/DC converter to provide traction energy or absorb regenerative braking energy. In the Boost and Buck modes, the state-space averaging method is used to establish a model and perform exact linearization. An adaptive sliding mode controller is designed, and simulation results show that it can effectively solve the problems of low energy utilization and large voltage fluctuations in urban rail electricity grids, and maximise the recovery and utilization of regenerative braking energy.

Event-triggered neuro sliding mode control for an uncertain Euler–Lagrange system: An impulsive dynamical system approach

This paper investigates the design of a neural network-based sliding mode control (SMC) with event-based sampling for an uncertain Euler–Lagrange (EL) system. The EL system considered here is connected to the controller over a shared communication network. The event-based SMC is designed to ensure the reduced usage of the communication and computational resources with robust tracking performance. The controller is designed under the assumption that the EL system dynamics are unknown; thus, a radial basis function neural network (RBFNN) is used to estimate it. The discrete weight update dynamics of RBFNN is formulated for a continuous-time EL system which makes the closed-loop system hybrid. Hence, the stability of the closed-loop system is investigated using the theory of impulsive dynamical systems with the admissible inter sampling time. Finally, a numerical example of an EL system is simulated to establish the effectiveness of the proposed control strategy.

Autonomous obstacle avoidance maneuvering of thrust-vectored airship with neural network control

There has been a renewed interest in recent times in airship technology owing to its potential usage for applications ranging from defense, scientific exploration, advertising to even remote monitoring. For airships to expand operational profile, further enhancement of configurational features and control development for full autonomy are key technologies gaining attention. In this paper, beginning with the mathematical modeling of a thrust-vectored airship, the integrated motion planning and controller development for vehicle autonomy, taking into account various uncertainties, are dealt with. A rapidly exploring random tree-based obstacle avoidance path planning exercise is carried out to chart out a trajectory in the presence of obstacles. Then, a neural network-based sliding mode controller is subsequently designed that learns the unknown equivalent control in sliding mode control framework to track the reference trajectory. Simulation results presented at the end demonstrate the effectiveness of the approach.

Neural network-based sliding mode control for atmospheric-actuated spacecraft formation using switching strategy

This paper presents an adaptive neural networks-based control method for spacecraft formation with coupled translational and rotational dynamics using only aerodynamic forces. It is assumed that each spacecraft is equipped with several large flat plates. A coupled orbit-attitude dynamic model is considered based on the specific configuration of atmospheric-based actuators. For this model, a neural network-based adaptive sliding mode controller is implemented, accounting for system uncertainties and external perturbations. To avoid invalidation of the neural networks destroying stability of the system, a switching control strategy is proposed which combines an adaptive neural networks controller dominating in its active region and an adaptive sliding mode controller outside the neural active region. An optimal process is developed to determine the control commands for the plates system. The stability of the closed-loop system is proved by a Lyapunov-based method. Comparative results through numerical simulations illustrate the effectiveness of executing attitude control while maintaining the relative motion, and higher control accuracy can be achieved by using the proposed neural-based switching control scheme than using only adaptive sliding mode controller.

Neural Network-based Sliding Mode Control for Permanent Magnet Synchronous Motor

In this paper, a scheme of designing neural network-based sliding mode controller is proposed for dealing with the chattering phenomenon existing in conventional sliding mode controller because of its discrete control law. Firstly, the sliding mode control law is designed using equivalent control technology. Then the neural network and adaptive control are uesd to delete the chattering of sliding mode controller. The stability of control system is analyzed by Lyapunov sta- bility theory finally. Simulations and experiments demonstrate that the proposed neural network-based sliding mode con- troller not only achieves better control performance than the conventional sliding mode control system, but also is robust with regard to system parameter variation and external disturbance.

Neural Network-based Sliding Mode Control for Permanent Magnet Synchronous Motor

Neural Network-based Sliding Mode Control for Permanent Magnet Synchronous Motor