A diagonal recurrent neural network-based hybrid direct adaptiveSPSA control system

A direct adaptive simultaneous perturbation stochastic approximation (DA SPSA) control system with a diagonal recurrent neural network (DRNN) controller is proposed. The DA SPSA control system with DRNN has simpler architecture and parameter vector size that is smaller than a feedforward neural network (FNN) controller. The simulation results show that it has a faster convergence rate than FNN controller. It results in a steady-state error and is sensitive to SPSA coefficients and termination condition. For trajectory control purpose, a hybrid control system scheme with a conventional PID controller is proposed     

GA-based fuzzy reinforcement learning for control of a magneticbearing system

This paper proposes a TD (temporal difference) and GA (genetic algorithm)-based reinforcement (TDGAR) learning method and applies it to the control of a real magnetic bearing system. The TDGAR learning scheme is a new hybrid GA, which integrates the TD prediction method and the GA to perform the reinforcement learning task. The TDGAR learning system is composed of two integrated feedforward networks. One neural network acts as a critic network to guide the learning of the other network (the action network) which determines the outputs (actions) of the TDGAR learning system. The action network can be a normal neural network or a neural fuzzy network. Using the TD prediction method, the critic network can predict the external reinforcement signal and provide a more informative internal reinforcement signal to the action network. The action network uses the GA to adapt itself according to the internal reinforcement signal. The key concept of the TDGAR learning scheme is to formulate the internal reinforcement signal as the fitness function for the GA such that the GA can evaluate the candidate solutions (chromosomes) regularly, even during periods without external feedback from the environment. This enables the GA to proceed to new generations regularly without waiting for the arrival of the external reinforcement signal. This can usually accelerate the GA learning since a reinforcement signal may only be available at a time long after a sequence of actions has occurred in the reinforcement learning problem. The proposed TDGAR learning system has been used to control an active magnetic bearing (AMB) system in practice. A systematic design procedure is developed to achieve successful integration of all the subsystems including magnetic suspension, mechanical structure, and controller training. The results show that the TDGAR learning scheme can successfully find a neural controller or a neural fuzzy controller for a self-designed magnetic bearing system.     

基于神经网络监督控制的拥塞控制算法研究

提出了一个基于神经网络控制的主动队列管理（AQM）算法;研究了TCP/AQM拥塞控制系统的可逆性,并利用一种神经网络监督控制结构进行了AQM算法的设计。算法由一个三层前馈结构的神经网络控制器（neural network controller,NNC）和一个反馈控制器（feedback controller,FC）组成。NNC作为一个前馈控制器,通过FC产生的教师信号进行学习,以建立被控对象的逆动力学模型。仿真结果表明,提出的算法与PI（proportional-integral）算法相比,无论在瞬态性能     

Neural network controller with flexible structure based on feedback-error-learning approach

In practice, the back-propagation algorithm often runs very slowly, and the question naturally arises as to whether there are necessarily intrinsic computation and difficulties with training neural networks, or better training algorithms might exist. Two important issues will be investigated in this framework. One establishes a flexible structure, to construct very simple neural network for multi-input/output systems. The other issue is how to obtain the learning algorthm to achieve good performance in the training phase. In this paper, the feedforward neural network with flexible bipolar sigmoid functions (FBSFs) are investigated to learn the inverse model of the system. The FBSF has changeable shape by changing the values of its parameter according to the desired trajectory or the teaching signal. The proposed neural network is trained to learn the inverse dynamic model by using back-propagation learning algorithms. In these learning algorithms, not only the connection weights but also the sigmoid function parameters (SFPs) are adjustable. The feedback-error-learning is used as a learning method for the feedforward controller. In this case, the output of a feedback controller is fed to the neural network model. The suggested method is applied to a two-link robotic manipulator control system which is configured as a direct controller for the system to demonstrate the capability of our scheme. Also, the advantages of the proposed structure over other traditional neural network structures are discussed.     

CPBUM neural networks for modeling with outliers and noise

In this study, CPBUM neural networks with annealing robust learning algorithm (ARLA) are proposed to improve the problems of conventional neural networks for modeling with outliers and noise. In general, the obtained training data in the real applications maybe contain the outliers and noise. Although the CPBUM neural networks have fast convergent speed, these are difficult to deal with outliers and noise. Hence, the robust property must be enhanced for the CPBUM neural networks. Additionally, the ARLA can be overcome the problems of initialization and cut-off points in the traditional robust learning algorithm and deal with the model with outliers and noise. In this study, the ARLA is used as the learning algorithm to adjust the weights of the CPBUM neural networks. It tunes out that the CPBUM neural networks with the ARLA have fast convergent speed and robust against outliers and noise than the conventional neural networks with robust mechanism. Simulation results are provided to show the validity and applicability of the proposed neural networks.     

Compound control of a compound cosine function neural network and PD for manipulators

In this paper, a compound cosine function neural network controller for manipulators is presented based on the combination of a cosine function and a unipolar sigmoid function. The compound control scheme based on a proportional-differential (PD) feedback control plus the cosine function neural network feedforward control is used for the tracking control of manipulators. The advantages of the compound control are that the system model does not need to be identified beforehand in the manipulator control system and it can achieve better adaptive control in an on-line continuous learning manner. The simulation results for the two-link manipulator show that the proposed compound control has higher tracking accuracy and better robustness than the conventional PD controllers in the position trajectory tracking control for the manipulator. Therefore, the compound cosine function neural network controller provides a novel approach for the manipulator control with uncertain nonlinear problems.     

Predictive neuro-control of uncertain systems: design and use of a neuro-optimizer

We consider the problem of predictive control of uncertain stochastic discrete I/O systems. Given a model identification procedure able to give accurate output system estimates, e.g. a neural network approximation, we use another feedforward neural network to generate at each time step a constrained optimal control. Dynamic backpropagation is used to improve when necessary the controller network parameters. Both system and controller neural structures are first selected off-line by a statistical Bayesian procedure in order to make the predictive control minimizing process more efficient. The issue of stochastic stability of the closed-loop is considered. We developed this approach for the tracking control of such uncertain systems as biotechnological processes. Actual and simulated predictive neuro-control case studies in this field of application are proposed as illustrations. A comparison with a more classic quasi-Newton-based approach is also proposed, showing the interest of this neuro-control approach.     

Model predictive control of an industrial packed bed reactor using neural networks

This paper discusses an industrial application of a multivariable nonlinear feedforward/feedback model predictive control where the model is given by a dynamic neural network. A multi-pass packed bed reactor temperature profile is modelled via recurrent neural networks using the backpropagation through time training algorithm. This model is then used in conjunction with an optimizer to build a nonlinear model predictive controller. Results show that, compared with conventional control schemes, the neural network model based controller can achieve tighter temperature control for disturbance rejection     

Neural-network hybrid control for antilock braking systems

The antilock braking systems are designed to maximize wheel traction by preventing the wheels from locking during braking, while also maintaining adequate vehicle steerability; however, the performance is often degraded under harsh road conditions. In this paper, a hybrid control system with a recurrent neural network (RNN) observer is developed for antilock braking systems. This hybrid control system is comprised of an ideal controller and a compensation controller. The ideal controller, containing an RNN uncertainty observer, is the principal controller; and the compensation controller is a compensator for the difference between the system uncertainty and the estimated uncertainty. Since for dynamic response the RNN has capabilities superior to the feedforward NN, it is utilized for the uncertainty observer. The Taylor linearization technique is employed to increase the learning ability of the RNN. In addition, the on-line parameter adaptation laws are derived based on a Lyapunov function, so the stability of the system can be guaranteed. Simulations are performed to demonstrate the effectiveness of the proposed NN hybrid control system for antilock braking control under various road conditions.     

Variable structure controller with neural compensator

This paper proposes a new variable structure controller combined with a multilayer neural network using an error back-propagation learning algorithm. The neural network acts as a compensator for a conventional variable structure controller in order to improve the control performance when the initial assumptions of the uncertainty bounds of the system parameters are violated. Also, the proposed controller can reduce the steady-state error of a conventional variable structure controller using the boundary layer technique. Computer simulation results show that the proposed method is effective in controlling dynamic systems with unexpected large uncertainties.     

基于模糊神经网络PID控制器的EEC数字化设计

为实现航空发动机模拟式电子控制器(EEC)的数字化设计,以其低压压气机导流叶片调节通道为主要研究对象,提出一种模糊神经网络PID控制器,将模糊控制、神经网络、PID控制相结合,利用模糊控制专家经验优势和神经网络的自学习、自适应能力,优化PID控制参数,实现控制性能提升。仿真结果显示,基于模糊神经网络的PID控制器控制性能有较大提高,具有比常规神经网络PID控制器更小的超调量和更好的抗干扰性;适用于定常系统和非定常系统,具有更好的自适应性与鲁棒性;可应用于航空发动机模拟式电子控制器(EEC)的数字化设计。     

Direct adaptive neural control of nonlinear systems with extreme learning machine

A direct adaptive neural control scheme for a class of nonlinear systems is presented in the paper. The proposed control scheme incorporates a neural controller and a sliding mode controller. The neural controller is constructed based on the approximation capability of the single-hidden layer feedforward network (SLFN). The sliding mode controller is built to compensate for the modeling error of SLFN and system uncertainties. In the designed neural controller, its hidden node parameters are modified using the recently proposed neural algorithm named extreme learning machine (ELM), where they are assigned random values. However, different from the original ELM algorithm, the output weight is updated based on the Lyapunov synthesis approach to guarantee the stability of the overall control system. The proposed adaptive neural controller is finally applied to control the inverted pendulum system with two different reference trajectories. The simulation results demonstrate good tracking performance of the proposed control scheme.     

Adaptive recurrent neural network control using a structure adaptation algorithm

This paper proposes an adaptive recurrent neural network control (ARNNC) system with structure adaptation algorithm for the uncertain nonlinear systems. The developed ARNNC system is composed of a neural controller and a robust controller. The neural controller which uses a self-structuring recurrent neural network (SRNN) is the principal controller, and the robust controller is designed to achieve L ² tracking performance with desired attenuation level. The SRNN approximator is used to online estimate an ideal tracking controller with the online structuring and parameter learning algorithms. The structure learning possesses the ability of both adding and pruning hidden neurons, and the parameter learning adjusts the interconnection weights of neural network to achieve favorable approximation performance. And, by the L ² control design technique, the worst effect of approximation error on the tracking error can be attenuated to be less or equal to a specified level. Finally, the proposed ARNNC system with structure adaptation algorithm is applied to control two nonlinear dynamic systems. Simulation results prove that the proposed ARNNC system with structure adaptation algorithm can achieve favorable tracking performance even unknown the control system dynamics function.     

机器人视觉伺服中CMAC学习控制系统研究

根据小脑模型关联控制器（CMAC）收敛速度快,适于实时控制系统的特点,设计了一种基于CMAC学习控制方法的机器人视觉伺服系统。在该系统中,CMAC被用作前馈视觉控制器对常规反馈控制器进行补偿。所提出的CMAC控制器替代图像雅可比矩阵来获得目标图像特征和机器人关节运动之间2D/3D变换关系,通过其在线学习,可以使系统对摄像机标定误差不敏感,从而提高系统的鲁棒性。实验证明了所设计控制系统的有效性。     

The Chebyshev-polynomials-based unified model neural networks forfunction approximation

In this paper, we propose the approximate transformable technique, which includes the direct transformation and indirect transformation, to obtain a Chebyshev-Polynomials-Based (CPB) unified model neural networks for feedforward/recurrent neural networks via Chebyshev polynomials approximation. Based on this approximate transformable technique, we have derived the relationship between the single-layer neural networks and multilayer perceptron neural networks. It is shown that the CPB unified model neural networks can be represented as a functional link networks that are based on Chebyshev polynomials, and those networks use the recursive least square method with forgetting factor as learning algorithm. It turns out that the CPB unified model neural networks not only has the same capability of universal approximator, but also has faster learning speed than conventional feedforward/recurrent neural networks. Furthermore, we have also derived the condition such that the unified model generating by Chebyshev polynomials is optimal in the sense of error least square approximation in the single variable ease. Computer simulations show that the proposed method does have the capability of universal approximator in some functional approximation with considerable reduction in learning time.     

溶解氧浓度的前馈神经网络建模控制方法

针对污水处理过程溶解氧(DO)浓度控制问题,提出了一种基于前馈神经网络的建模控制方法(FNNMC).本文构造了神经网络建模控制系统,通过对建模神经网络和控制神经网络隐含层学习率的分析,证明了学习算法的收敛性以及整个系统的稳定性.最后,本文基于国际基准的Benchmark Simulation Model No.1 (BSMl)进行了仿真实验,验证了合理选取学习率的重要性,并通过与PID和模型预测控制(MPC)等已有控制方法的比较,验证了神经网络建模控制方法针对污水处理过程溶解氧浓度控制具有良好的建模能力,更高的控制精度以及更好的动态响应能力.     

Synthesis of the sliding-mode neural network controller for unknown nonlinear discrete-time systems

This paper develops a sliding-mode neural network controller for a class of unknown nonlinear discrete-time systems using a recurrent neural network (RNN). The control scheme is based on a linearized expression of the nonlinear system using a linear neural network (LNN). The control law is proposed according to the discrete L yapunov theory. With a modified real-time recurrent learning algorithm, the RNN as an estimator is used to estimate the unknown part in the control law in on-line fashion. The stability of the control system is guaranteed owing to the on-line learning ability of the RNN algorithm. The proposed control scheme is applied to numerical problems and simulation results that it is very effective.     

Neural network based model predictive control for a steel pickling process

A multi-layer feedforward neural network model based predictive control scheme is developed for a multivariable nonlinear steel pickling process in this paper. In the acid baths three variables under controlled are the hydrochloric acid concentrations. The baths exhibit the normal features of an industrial system such as nonlinear dynamics and multi-effects among variables. In the modeling, multiple input, single-output recurrent neural network subsystem models are developed using input–output data sets obtaining from mathematical model simulation. The Levenberg–Marquardt algorithm is used to train the process models. In the control (MPC) algorithm, the feedforward neural network models are used to predict the state variables over a prediction horizon within the model predictive control algorithm for searching the optimal control actions via sequential quadratic programming. The proposed algorithm is tested for control of a steel pickling process in several cases in simulation such as for set point tracking, disturbance, model mismatch and presence of noise. The results for the neural network model predictive control (NNMPC) overall show better performance in the control of the system over the conventional PI controller in all cases.     

高速公路匝道回归神经网络控制器的设计

提出用回归神经网络进行入口匝道控制的思路。阐述了Elman回归神经网络原理与入口匝道控制原理,选取上、下游时间占有率和车速作为匝道控制器的输入量,并设计了Elman回归神经网络入口匝道控制器,采用一种改进的算法对回归神经网络进行训练。仿真实验表明,该控制器学习误差小,泛化能力好,具有良好的应用前景。     

Adaptive RTRL based neurocontroller for damping subsynchronous oscillations using TCSC

Modern interconnected electrical power systems are complex and require perfect planning, design and operation. Hence the recent trends towards restructuring and deregulation of electric power supply has put great emphasis on the system operation and control. Flexible AC transmission system (FACTS) devices such as thyristor controlled series capacitor (TCSC) are capable of controlling power flow, improving transient stability and mitigating subsynchronous resonance (SSR). In this paper an adaptive neurocontroller is designed for controlling the firing angle of TCSC to damp subsynchronous oscillations. This control scheme is suitable for non-linear system control, where the exact linearised mathematical model of the system is not required. The proposed controller design is based on real time recurrent learning (RTRL) algorithm in which the neural network (NN) is trained in real time. This control scheme requires two sets of neural networks. The first set is a recurrent neural network (RNN) which is a fully connected dynamic neural network with all the system outputs fed back to the input through a delay. This neural network acts as a neuroidentifier to provide a dynamic model of the system to evaluate and update the weights connected to the neurons. The second set of neural network is the neurocontroller which is used to generate the required control signals to the thyristors in TCSC. This is a single layer neural network. Performance of the system with proposed neurocontroller is compared with two linearised controllers, a conventional controller and with a discrete linear quadratic Gaussian (DLQG) compensator which is an optimal controller. The linear controllers are designed based on a linearised model of the IEEE first benchmark system for SSR studies in which a modular high bandwidth (six-samples per cycle) linear time-invariant discrete model of TCSC is interfaced with the rest of the system. In the proposed controller, since the response time is highly dependent on the number of states of the system, it is often desirable to approximate the system by its reduced model. By using standard Hankels norm approximation technique, the system order is reduced from 27 to 11th order by retaining the dominant dynamic characteristics of the system. To validate the proposed controller, computer simulation using MATLAB is performed and the simulation studies show that this controller can provide simultaneous damping of swing mode as well as torsional mode oscillations, which is difficult with a conventional controller. Moreover the fast response of the system can be used for real-time applications. The performance of the controller is tested for different operating conditions.