基于神经网络的机器人运动模型辨识及其实验研究

针对机器人楚模中不确定因素的影响，采用神经网络辨识机器人输入输出间的非线性关系，建立机器人的运动学模型，为了提高神经网络的辨识速度，基于Elman动态递归网络，通过增加网络输入输出的部分信息，提出一种新的动态神经网络结构——状态廷迟输入动态递归神经网络(SDIDRNN)，提高了网络的学习速度和稳态精度。以PowerCube^TM模块化机器人为研究对象，把根据机器人返回的关节位置信息和利用OPTOTRAK3020三维运动测量系统测得的机器人末端位置信患作为SDIDRNN的学习样本，对包含各种影响因素的机器人运动模型进行辨识，得到了满意的结果，说明了该神经网络的优越性。     

神经网络并联辨识算法的收敛性研究

神经网络可用来建立非线性动态系统的模型，其辨识模型可分为串联并联辨识模型和并联辨识模型两种，后者的思路源于基于参考模型自适应方案的输出误差辨识模型，对观测扰动有较强的抑制能力。本文对这种神经网络并联辨识结构的收敛性进行了研究，指出在网络参数满足一定条件时并联预测过程收敛，且并联辨识算法具有局部收敛性，仿真实验验证了上述结论。     

基于动态递归神经网络的木材干燥模型辩识

木材干燥是一个复杂的非线性系统，因此利用传统的系统辨识方法难以建立其准确的模型。本文利用动态递归神经网络的特点，提出了基于动态递归神经网络的木材干燥模型辨识方法，仿真结果表明，利用动态递归神经网络所建立的模型是有效的。     

一类非线性系统LMF优化迭代神经网络控制器

提出一类非线性系统ＬＭＦ优化迭代神经网络控制器的设计方法。该方法在正向神经网络辨识模型的基础上，应用ＬＭＦ优化迭代方法进行控制器设计，理论证明，只要神经网络辨识模型的精度足够高，就会获得很好的控制精度。为补偿辨识和迭代学习误差，给出了通过引入反馈补偿控制器提高控制精度的方法，仿真结果证明了该方法的有效性。     

清洗机器人自适应PID控制研究

对于清洗机器人这样一个复杂的非线性、时变系统,常规PID控制方法难以达到满意的控制效果.提出了一种BP神经网络与常规PID(比例、积分、微分)控制相结合的智能控制方法,利用辨识网络逼近被控系统获得BP神经网络学习所需梯度信息,从而实现PID控制参数的在线调整,以适应控制系统的动态变化.对系统辨识模型进行了正弦输入下的逼近仿真;在对建立的运动学模型经离散化处理,得到其传递函数的基础上,对控制系统进行的角阶跃输入响应进行了对比仿真分析.仿真结果表明,辨识模型能很好的逼近被控对象,基于BP神经网络的自适应PID控制方法在解决清洗机器人控制问题时,提高了控制的响应时效,增强了系统的稳定性,获得了比传统PID控制更好的控制品质.     

线性神经网络及在系统辨识中的初步应用

该文从基本的智能控制技术——神经网络(NN)技术出发，探讨了神经网络用于系统辨识与建模的基本理论，分析了线性系统神经网络建模的规律，提出了一种利用线性神经网络进行系统辨识的方法。该辨识方法显示出很强的处理问题的能力，无需辨别系统阶次，辨识结构简单，收敛速度快，仿真结果表明这种方法的有效性和可行性。该文共分为四部分，第一部分介绍了神经网络用于系统辨识的特征，第二部分讲述了线性神经网络的工作原理，包括线性神经网络的模型、传递函数、学习规则及训练过程，第三部分讲述了线性神经网络进行系统辨识的仿真实例，第四部分对上述内容作了简要小结。     

基于神经网络的动态系统逆模型辨识及闭环控制

本文提出一种动态线性或非线性系统的神经网络逆模型辨识结构，并引出两种ＰＩＤ与神经网络逆模型相结合的自适应控制方案，神经网络模型采用基于Ｕ－Ｄ分解卡尔曼滤波学习算法（ＵＤＫ）的动态前向多层网、仿真结果表明了所述辨识方案的有效性及特点 。     

一种工业机器手的自适应阻抗控制方法

提出一种基于任务空间的直接自适应阻抗方法，它不要求辨识机器手动态模型结构和参数，不需要计算机器手的运动学逆变换，因此，避免了基于机器手模型线性参数辨识的控制方法的缺点。     

一种基于改进神经网络的系统辨识方法

该文在分析神经网络辨识技术特点及现状的基础上,将BP神经网络结构和遗传算法相结合,设计了一种适用于非线性系统的辨识器模型。该辨识器模型首先建立初始的BP神经网络结构,再利用遗传算法对BP神经网络的权值和阈值进行优化,从而优化BP神经网络,通过迭代最终建立辨识器模型。最后,通过一个三阶非线性多输入单输出系统的仿真实验证明了所设计的辨识器具有辨识时间短、辨识精度高的特点,为神经网络辨识技术的研究提供了新的思路和方法。     

基于ANN逆模型补偿的复合控制系统仿真

为对复杂非线性系统进行辨识建模和实施有效控制，分析了基于神经网络的非线性系统逆模型的辨识和控制原理，研究了基于神经网络的非线性系统逆模型补偿的复合控制方法。基于复合控制思想，时常规PID控制器＋前馈神经网络逆模型补偿的复合控制结构方案进行了仿真。仿真结果表明，基于神经网络的非线性系统逆模型补偿的复合控制结构方案是有效的、相对简单的网络结构，可提高逆模型的泛化能力和非线性系统的控制精度。     

多智能体强化学习的机械臂运动控制决策研究

针对传统运动控算法存在环境适应性较差,效率低的问题。可以利用强化学习在环境中不断去探索试错,并通过奖励函数对神经网络参数进行调节的方法对机械臂的运动进行控制。但是在现实中无法提供机械臂试错的环境,采用Unity引擎平台来构建机械臂的数字孪生仿真环境,设置观察状态变量和设置奖励函数机制,并提出在该模型环境中对PPO(proximal policy optimization)与多智能体（agents）结合的M-PPO算法来加快训练速度,实现通过强化学习算法对机械臂进行智能运动控制,完成机械臂执行末端有效避障快速到达目标物体位置,并通过该算法与M-SAC（多智能体与Soft Actor-Critic结合）和PPO算法的实验结果进行分析,验证M-PPO算法在不同环境下机械臂运动控制决策调试上的有效性与先进性。实现孪生体自主规划决策,反向控制物理体同步运动的目的。     

捕获目标卫星后组合体航天器模糊神经网络滑模控制

探讨了漂浮基空间机械臂系统在轨捕获参数未知目标卫星后组合体航天器的镇定控制问题.首先在耦合空间机械臂系统捕获目标卫星操作过程动量、冲量的传递的基础上,建立了适用于漂浮基空间机械臂系统在轨捕获漂浮卫星控制系统设计的组合体航天器数学模型.利用该模型,设计了一种基于模糊高斯基神经网络的非奇异Terminal滑模控制算法.提出的控制算法不仅不要求系统动力学方程关于惯性参数呈线性函数关系,而且也不需要预知系统惯性参数;由于利用神经网络的自学习能力修正模糊控制的控制规则和隶属函数,这样在系统参数识别中,模糊神经网络可减少模糊规则数,更适应于空间机械臂系统在轨捕获的实际应用.最后通过仿真试验对比结果验证了所提出的控制算法的有效性.     

Identification and decentralized adaptive control using dynamicalneural networks with application to robotic manipulators

Efficient implementation of a neural network-based strategy for the online adaptive control of complex dynamical systems characterized by an interconnection of several subsystems (possibly nonlinear) centers on the rapidity of the convergence of the training scheme used for learning the system dynamics. For illustration, in order to achieve a satisfactory control of a multijointed robotic manipulator during the execution of high speed trajectory tracking tasks, the highly nonlinear and coupled dynamics together with the variations in the parameters necessitate a fast updating of the control actions. For facilitating this requirement, a multilayer neural network structure that includes dynamical nodes in the hidden layer is proposed, and a supervised learning scheme that employs a simple distributed updating rule is used for the online identification and decentralized adaptive control. Important characteristic features of the resulting control scheme are discussed and a quantitative evaluation of its performance in the above illustrative example is given.     

Learning from hint for the conservative motion of the constrained industrial redundant manipulators

A neural network (NN)-based kinematic inversion of industrial redundant arms is developed in this paper to conserve the joint configuration in cyclic trajectories. In the developed approach, the Widrow–Hoff NN with an online adaptive learning algorithm derived by applying Lyapunov approach is introduced. Since this kinematic inversion has an infinite number of joint angle vectors, a fuzzy neural network system is designed to provide an approximate value for that vector. Feeding this vector as an additional hint input vector to the NN limits and guides the output of the NN within the self-motion of the manipulator. The derivation of the candidate Lyapunov function, which is designed to achieve the joint configurations conservation in addition to the joint limits avoidance, leads to a computationally efficient online learning algorithm of the NN. Simulations are conducted for the PA-10 redundant manipulator to bear out the efficacy of the developed approach for tracking closed trajectories.     

Visual servoing of redundant manipulator with Jacobian matrix estimation using self-organizing map

Vision based redundant manipulator control with a neural network based learning strategy is discussed in this paper. The manipulator is visually controlled with stereo vision in an eye-to-hand configuration. A novel Kohonen’s self-organizing map (KSOM) based visual servoing scheme has been proposed for a redundant manipulator with 7 degrees of freedom (DOF). The inverse kinematic relationship of the manipulator is learned using a Kohonen’s self-organizing map. This learned map is shown to be an approximate estimate of the inverse Jacobian, which can then be used in conjunction with the proportional controller to achieve closed loop servoing in real-time. It is shown through Lyapunov stability analysis that the proposed learning based servoing scheme ensures global stability. A generalized weight update law is proposed for KSOM based inverse kinematic control, to resolve the redundancy during the learning phase. Unlike the existing visual servoing schemes, the proposed KSOM based scheme eliminates the computation of the pseudo-inverse of the Jacobian matrix in real-time. This makes the proposed algorithm computationally more efficient. The proposed scheme has been implemented on a 7 DOF PowerCube? robot manipulator with visual feedback from two cameras.     

带神经网络补偿的极点配置广义最小方差自校正控制

首先用一个常规线性模型对被控对象进行辨识,再对线性模型辨识的余差用一个神 经网络进行补偿．线性模型和神经网络共同构成对象的辨识模型,并基于这一模型提出了一 种显式极点配置广义最小方差自校正控制．该方法适用于非线性对象,且具有较高精度和较 快的收敛速度,具有较强的鲁棒性．     

Robust Interaction Control of a Mobile Manipulator – Dynamic Model Based Coordination

This paper discusses the modeling and control of a spatial mobile manipulator which consists of a robotic manipulator mounted upon a wheeled mobile platform. The nonholonomic model, which assumes perfect contact between the wheels and the ground, is obtained using the Lagrange–d'Alembert formulation. Also, the dynamic model, which considers slip of the platform's tires, is developed using the Newton–Euler method and incorporates Dugoff's tire friction model. The complexity of the model is increased by introducing kinematic redundancy which is created when a multi-linked manipulator is used. The kinematic redundancy is resolved by decomposing the mobile manipulator into two subsystems; the mobile platform and the manipulator. Based on the coordination scheme used to resolve the kinematic redundancy, a robust interaction control algorithm, in which suitable controllers are designed for the two subsystems, is developed and applied. The adverse effect of the wheel slip on the tracking of commanded motion is discussed in the simulation. For the dynamic model, a robust control approach is employed to minimize the harmful effect of the wheel slip on the tracking performance. Simulation results show the promise of the developed algorithm.     

果实摘取机械手的果实识别方法与摘取路径规划

通过图形分割处理技术从采集到的原始图形中分割出目标果实图形,再利用由遗传算法优化后的神经网络算法构建果实图形识别模型,完成对果实图形的识别。使用图像识别技术识别出机械手当前视线中的果实数量和位置,利用遗传算法对机械手的运行路径进行优化,并将其结果与随机路径规划和人工路径规划的结果进行对比。研究结果表明：随着样本中的完整果实比例逐渐减少,果实识别模型的识别准确率均有所下降,在仅有60%完整果实比例的样本中所研究的识别模型仍保持较高的识别准确率。使用遗传优化算法得到的机械手行进路径相比随机路径规划和人工路径规划所消耗的成本更低。并且随着果实数量的增加,遗传优化算法得到的机械手行进路径消耗的成本低的优势更加突出。     

Query based model learning and stable tracking of a robot arm using radial basis function network

In the context of a robot manipulator, a generalized neural emulator over the complete workspace is very difficult to obtain because of dimensionally insufficient training data. A query based learning algorithm is proposed in this paper that can generate new examples where control inputs are independent of states of the system. This algorithm is centered around the concept of network inversion using an extended Kalman filtering based algorithm. This is a novel idea since robot manipulator is an open loop unstable system and generation of control input independent of state is a research issue for neural model identification. Two trajectory independent stable control schemes have been designed using the neural emulator. One of the control schemes uses forward-inverse-modeling approach to update the controller parameters adaptively following Lyapunov function synthesis technique. The proposed scheme is trajectory independent unlike the back-propagation scheme. The second type of controller predicts the minimum variance estimate of control action using recall process (network inversion) and the control law is derived following a Lyapunov function synthesis approach so that the closed loop system consisting of controller and neural emulator remains stable. The simulation experiments show that the model validation approach is efficient and the proposed control schemes guarantee stable accurate tracking.     

一种结合PSO及改进BP网络的辨识方法

当辨识神经网络的类型和结构确定后,初始权值等辨识参数直接影响到辨识效果,而依靠先验知识试凑而得的参数值往往难以达到最佳效果。针对这一问题,提出了一种结合粒子群（PSO）算法及引入动量项的改进BP网络的辨识方法,利用PSO对改进BP网络辨识的初始权值/偏置、学习率、动量系数进行寻优,并将优化后的神经网络模型用在控制系统中进行修正,进一步完善辨识模型。应用在热工系统中,仿真结果表明了该辨识方法的有效性。