A REAL-TIME ADAPTIVE CONTROL ALGORITHM USING NEURAL NETS WITH PERTURBATION

ＮＴＲＯＤＵＣＴＩＯＮＡｆｔｅｒｍｕｃｈｒｅｓｅａｒｃｈａｎｄｓｏｍｅｓｕｃｃｅｓｓｆｕｌａｐ ｐｌｉｃａｔｉｏｎｓｏｆＡｒｔｉｆｉｃｉａｌＮｅｕｒａｌＮｅｔｗｏｒｋ (ＡＮＮ)ｔｏｃｏｎｔｒｏｌｓｙｓｔｅｍｓ,ＡＮＮｈａｓｂｅｃｏｍｅｏｎｅｏｆｔｈｅｉｍ ｐｏｒｔａｎｔｔｏｏｌｓｔｏｓｈｏｗｔｈｅｉｎｔｅｌｌｉｇｅｎｃｅｏｆｔｈｅｈｕ ｍａｎｂｒａｉｎ .ＳｉｎｃｅＡＮＮｃａｎｂｅｕｓｅｄｔｏｄｅｓｃｒｉｂｅａｎｙｎｏｎｌｉｎｅａｒｃｏｎｔｉｎｕｏｕｓｆｕｎｃｔｉｏｎ ,ｉｔｈａｓｄｒａｗｎｍｕｃｈａｔ…     

Elevator group-control policy based on neural network optimized by genetic algorithm

Aiming at the diversity and nonlinearity of the elevator system control target, an effective group method based on a hybrid algorithm of genetic algorithm and neural network is presented in this paper. The genetic algorithm is used to search the weight of the neural network. At the same time, the multi-objective-based evaluation function is adopted, in which there are three main indicators including the passenger waiting time, car passengers number and the number of stops. Different weights are given to meet the actual needs. The optimal values of the evaluation function are obtained, and the optimal dispatch control of the elevator group control system based on neural network is realized. By analyzing the running of the elevator group control system, all the processes and steps are presented. The validity of the hybrid algorithm is verified by the dynamic imitation performance.     

基于RBF神经网络的不确定机械臂自适应反演控制研究

针对机械臂位置跟踪问题,基于RBF神经网络控制理论,提出了一种自适应反演控制方法．该方法利用反演控制技术解决了系统的非线性问题,通过神经网络对系统中不确定函数进行逼近,实现神经网络自适应反演控制,最后通过Lyapunov稳定性理论证明所得闭环系统的最终一致有界．仿真结果证明该算法的有效性．     

Simultaneous perturbation stochastic approximation algorithm combined with neural network and fuzzy simulation

In order to solve three kinds of fuzzy programm model, fuzzy chance-constrained programming mode ng models, i.e. fuzzy expected value and fuzzy dependent-chance programming model, a simultaneous perturbation stochastic approximation algorithm is proposed by integrating neural network with fuzzy simulation. At first, fuzzy simulation is used to generate a set of input-output data. Then a neural network is trained according to the set. Finally, the trained neural network is embedded in simultaneous perturbation stochastic approximation algorithm. Simultaneous perturbation stochastic approximation algorithm is used to search the optimal solution. Two numerical examples are presented to illustrate the effectiveness of the proposed algorithm.     

高层建筑群控电梯应急疏散调度算法的研究

高层建筑的电梯不仅能够用于正常的客流运输,而且也能够在应急情况下用于紧急疏散.文章将电梯的应急疏散作用纳入电梯群控系统,提出静态分区与基于模糊控制的多目标优化算法结合的调度方法.介绍了高层建筑应急疏散的特点以及算法的思想,并通过仿真试验验证了算法的有效性.     

工业监测控制系统中的实时数据库研究和设计

本文针对常规关系型数据库系统在工业监测控制实时应用中存在的缺陷，提出了将实时技术和传统的数据库技术相结合的方法，通过实例分析了一个实用的监控系统中实时数据库系统的构成与分类，提出了监控系统的实时数据库模型，并对其体系结构、系统特色、C／S访问控制、I/O访问控制、分布式机制等方面进行了探讨。     

Construction simulation and real-time control for high arch dam

A method of combining dynamic simulation with real-time control was proposed to fit the randomness and uncertainty in the high arch dam construction process. The mathematical logic model of high arch dam construction process was established. By combining dynamic construction simulation with schedule analysis, the process of construction schedule forecasting and analysis based on dynamic simulation was studied. The process of real-time schedule control was constructed and some measures for dynamic adjustment and control of construction schedule were provided. A system developed with the method is utilized in a being constructed hydroelectric project located at the Yellow River in northwest China, which can make the pouring plan of the dam in the next stage (a month, quarter or year) to guide the practical construction. The application result shows that the system provides an effective technical support for the construction and management of the dam. ZHONG Denghua, born in 1963, male, Dr, Prof. Supported by National Natural Science Foundation of China(No.50539120), National Basic Research Program of China(“973”Program,No. 2007 CB714101), National Science Fund for Distinguished Young Scholars of China(No.50525927), and National Natural Science Foundation of China (No.50579045).     

基于神经网络的超精密定位控制

针对精密定位装置存在非线性,精确数学模型难于建立的缺陷,提出了精密定位的神经网络控制方法.将BP神经网络应用于该控制系统中,系统以光栅常数100μm的光栅为定位标记,以激光衍射产生的莫尔光光强及光强的变化率为神经网络的输入变量,利用神经网络的自学习功能进行精密定位控制.建立了精密定位的神经网络控制模型,模型由输入层、隐层和输出层3层神经元组成,通过对光强及光强变化率的映射,得到电机驱动信号.实验结果表明,使用神经网络控制,控制响应快,稳定性好,鲁棒性强,可有效改善控制质量,提高定位速度,系统可获得±0·5μm的定位精度.     

Parameters optimization for exponentially weighted moving average control chart using generalized regression neural network

1 Introduction a The statistical process monitoring and quality control in the industry manufacturing using exponentially weighted moving average (EWMA) control chart have attracted much attention recently. Parameter selection is the key to optimal design…     

基于倒立摆系统前馈型神经网络实验的开发

介绍一个基于MATLAB/SIMULINK和直线一级倒立摆系统的前馈型人工神经网络的实验方案。设计了一级倒立摆系统BP神经网络控制器,并完成了仿真和实物控制实验。实验结果有利于对前馈型人工神经网络的本质认识。     

基于神经网络的数字广播信号无线定位算法(英文)

为了提高非视距(NLOS)环境下无线定位的准确性和可靠性,提出了一种利用数字广播信号进行移动台定位的神经网络方法.该方法利用神经网络的学习特性和逼近任意非线性函数的能力,建立到达时间(TOA)和到达时间差(TDOA)测量数据与坐标之间的映射关系.将神经网络的连接权值作为非线性动态系统的状态量进行估计,用基于扩展卡尔曼(EKF)的实时神经网络训练算法来训练多层感知器网络.由于基于EKF的训练算法给出的是连接权值的近似最小方差估计,其收敛性要优于误差反向传播(BP)算法.仿真结果表明,该算法在NLOS环境下有较高的定位精度,性能优于BP基的神经网络算法和最小二乘算法;且该定位方法不依赖于特定的NLOS误差分布,也无需视距(LOS)和非视距识别.     

基于WinCC的水箱液位控制系统实时仿真平台开发北大核心

该文以耦联双容水箱为例,介绍了一种基于WinCC的水箱液位控制系统实时仿真平台开发。首先用WinCC构建双容水箱工艺流程动态监控画面;然后在C脚本下实现对水箱液位非线性数学模型的求解和增量PID控制量的计算;最后通过设置相同的采样时间并关联相应的变量,实现监控动态画面和C脚本之间的通信,组建成一个完整的闭环控制系统,并进行手动/自动模式的测试。该仿真平台能够用动画实时地展示水箱液位变化过程,立体感和真实感强,可实现在线修改控制器参数、手动/自动两种模式无扰切换等功能,人机交互效果好,是一个有效的验证各类控制算法及其参数设定的仿真实验平台。     

ReinforcementBased Fuzzy Neural Network Control with Automatic Rule Generation

１　Ｉｎｔｒｏｄｕｃｔｉｏｎ　Ａｆｕｚｚｙｎｅｕｒａｌｎｅｔｗｏｒｋｃｏｎｔｒｏｌｌｅｒｂａｓｅｄｏｎｔｈｅｃｏｍｂｉｎａｔｉｏｎｏｆｆｕｚｚｙｌｏｇｉｃｓｙｓｔｅｍａｎｄｎｅｕｒａｌｎｅｔｗｏｒｋｈａｓｕｎｉｑｕｅｃｈａｒａｃｔｅｒｉｓｔｉｃｓ．Ｉｔｎｏｔｏｎｌｙｈａｓｔｈｅｆｕｎｃｔｉｏｎｏｆｌｅａｒｎｉｎｇａｎｄａｓｓｏｃｉａｔｉｖｅｍｅｍｏｒｙｂｕｔａｌｓｏｃｌｅａｒｌｙｒｅｐｒｅｓｅｎｔｓａｋｉｎｄｏｆｓｔｒｕｃｔｕｒｅｄｋｎｏｗｌｅｄｇｅａｎｄｄｏｅｓｆｕｚｚｙｉｎｆｅｒｅｎ…     

基于RBF神经网络辨识的直接甲醇燃料电池电堆非成性建模与自适应模糊控制

1 Introduction Direct methanol fuel cell ( DMFC) is desirable toserve as the power systemfor portable devices such ascellular phones , portable computers ,etc. due to thetheoretically high energy density and the liquid fuelused that can be stored and tran…     

单神经元自适应PID在漂白温度控制中的应用

针对纸浆漂白过程中漂白塔温度大滞后、结构参数时变的特点,提出基于单神经元的自适应PID来控制该对象,使得对象在其结构参数变化的情况下在线实时调整PID的三个参数,从而提高了对象的鲁棒性和精确性.实践结果表明,该控制器具有算法简单、易实现和适应性强等特点.     

Nonlinear model predictive control with guaranteed stability based on pseudolinear neural networks

~~identificationis7-7-1,where3nm==,andtheRPEalgorithmisusedtoupdatetheweightingofPNN.Thewholetrainingprocessuses800iterations.InordertoovercometheinaccuracyofPNNmodel,thecontrollerstructureisacompositeoneasfbff()()()ututut=+,(37)wherefb()utistheoutputoffeedbackcontroller,ff()utistheoutputofpredictivecontrollerdescribedbyEq.(14),with0.20=,0.40=,andmax5K=.Insimulatedclosedloopcontrol,ufb(t)isaproportionalcontroller,fb()()Putket=and5.0Pk=.Theset-pointofthesystemisd0.15,if040,and120()0.24,if4…     

A filter algorithm for multi-measurement nonlinear system with parameter perturbation

INTRODUCTION For the filtering problem of multi-measurement nonlinear system with parameter perturbation, we need to consider the following aspects. First, although the parameter uncertainty does not change the system configuration, it makes the filter designed based on a fixed parameter unsuitable for an alterable parameter. Second, there are false alarms in the raw measure- ments which are received by multiple sensors. Third, since the optimal nonlinear Bayesian estimation is impossible…     

一种可自适应优化模糊控制规则的控制系统

提出一种模糊控制规则可自适应调整的控制算法.该算法利用模糊逻辑系统构建模糊辨识器和模糊控制器,依据辨识器在线预估输出误差与模糊控制规则相关参数优化之间的内在联系及自适应优化模糊控制规则.仿真实验表明了该算法的有效性.     

Nonlinear model predictive control with guaraneed stability based on pesudolinear neural networks

A nonlinear model predictive control problem based on pseudo-linear neural network (PNN) is discussed, in which the second order on-line optimization method is adopted. The recursive computation of Jacobian matrix is investigated. The stability of the closed loop model predictive control system is analyzed based on Lyapunov theory to obtain the sufficient condition for the asymptotical stability of the neural predictive control system. A simulation was carried out for an exothermic first-order reaction in a continuous stirred tank reactor. It is demonstrated that the proposed control strategy is applicable to some of nonlinear systems.     

基于模糊神经网络PID的无刷直流电机控制系统研究

针对传统PID控制在无刷直流电机控制系统中达不到良好的控制效果的问题,在无刷直流电机的数学模型基础之上,设计一种模糊神经网络自适应PID控制器,该控制器利用模糊控制非线性控制作用和BP神经网络的学习能力及适应能力相结合对PID参数进行在线实时调整。对基于模糊神经网络自适应PID控制器的无刷直流电机的双闭环控制系统进行仿真实验,实验结果表明,可以提高控制系统的响应速度,减小超调量,对负载及电机参数的变化都有较强的鲁棒性。