Position domain contour control for multi-DOF robotic system

Contour tracking control is one of the fundamental operations for robotic systems. In this paper, a position domain PD control is developed to control a multi-DOF nonlinear robotic system for improving contour tracking performance. In this new position domain control system, a robotic system is viewed as a master–slave system where the master motion is used as an independent reference through equidistantly sampling, while slave motions are described as functions of the master motion according to contour tracking requirements. A position domain dynamic model of the robotic system based on the master motion is developed through one-to-one mapping of the original dynamic model from time domain to position domain. Stability analysis is conducted for the proposed position domain PD control, the global boundedness of the tracking errors is guaranteed through the Lyapunov method, and the effectiveness is successfully verified through simulation study for linear and nonlinear contour tracking problems. Compared results demonstrate that the position domain PD control is better than its time domain counterpart for contour tracking of multi-DOF robotic systems.     

A model predictive control approach for time optimal point-to-point motion control

This paper presents a new model predictive control method for time-optimal point-to-point motion control of mechatronic systems. The formulation of time-optimal behavior within the model predictive control framework and the structure of the underlying optimization problem are discussed and modifications are presented in order to decrease the computational load of the numerical solution method such that sampling rates in the millisecond range and long prediction horizons for large point-to-point motions are feasible. An extensive experimental validation on a linear motor drive and an overhead crane setup demonstrates the advantages of the developed time-optimal model predictive control approach in comparison with traditional model predictive control.     

Cross-coupling position command shaping control in a multi-axis motion system

A new structure of a cross-coupling position command shaping controller (CPCSC) for precise tracking in multi-axis motion control is proposed in this paper. This controller feedforwards the cross-coupling terms, based on the geometrical relationship between the tracking and contouring errors, to compensate for the contouring error in real-time. Compared with the conventional multi-axis cross-coupling control (CCC) system, this new structure has the advantage that its compensators in CCC have a simpler design process than conventional ones, as does its stability analysis. The proposed controller is evaluated and compared experimentally with a traditional uncoupled and a conventional CCC controller on a multi-axis positioning system controlled by microcomputer. The experimental results show that the new structure remarkably reduces contour error. In addition, this new controller can be implemented easily on most current systems by reprogramming the reference position command subroutine.     

多台水泵并联的最优化方法

在城市污水处理系统中,由于来水流量的变化,往往需要开启多台水泵,传统的控制方法是基于专业人士依据自身经验对其进行控制,精确度不够,而且耗能巨大。通过对水泵各参数之间的关系进行较为系统的分析,建立水泵系统的数学模型,依据最优控制的方法,求出轴功率的最小值,并以此确定了开启水泵的台数、频率和流量等参数,实现了耗能最低。上述控制方法不仅适用于同一型号水泵并联,同时也为不同型号水泵并联提供了依据,达到了节能的目的。     

最优滑模飞行控制系统的设计与仿真

针对机动飞机按时标分离原则可分为快慢两个子系统而形成两环控制结构的特点，提出了最优滑模飞控系统的设计。外环控制器的设计采用基于依赖状态的Riccati方程最优控制器，用以产生最优滑模面，以保证整个飞控系统具有一定的性能鲁棒。内环控制器设计时采用滑动模控制以减小飞控系统对参数变化、模型误差、外部干扰敏感，具有一定的稳定鲁棒性。最后对机动飞机作大机动仿真，仿真结果表明该飞控系统是有效的。     

基于视觉皮层机制的轮廓整合的神经网络模型

本文以生物视觉功能及其机理有关的研究成果为主要依据,通过分析、建模和模拟计算等方式建立进化的图像目标轮廓组织整合的理论、方法和模型。首先通过分析初级视觉通道内（LGN—〉视觉皮层V1区—〉视觉皮层V2区神经元）目标轮廓的组织整合机理,建立具有广泛视觉心理学和生理学基础人工神经网络模型;然后将该模型应用于自然场景图像的目标轮廓组织及整合的模拟计算中,并根据实验结果对模型做出分析与评估。     

Hierarchical Learning-Guided human motion quality assessment in big data environment

Image information may be distorted during acquisition, processing, compression, and transmission. It is necessary to propose an intelligent image quality assessment model toward big data environment to quantify the degree of distortion of the image. This paper proposes a quality assessment model for human motion images. In complex scenes, the human body's action posture can be taken as an important feature point. Usually, in different scenes, the parts that affect the quality of the human body's posture are different. In other words, the weights of feature points that affect quality are different in different scenarios. However, due to the categorization of human movements, we can learn the quality assessment methods of different types of movements through sample training. Inspired by feature learning in the field of machine learning, we propose a hierarchical quality learning approach. We cast quality assessment as quality feature learning and layer by layer. The hierarchical quality learning method is based on deep reinforcement learning. The key part is that the method focuses on the region that containing more information on the features of the quality and enlarges the region layer by layer. Finally, we can determine the part of the body that affects the quality assessment. We compare this method with the subjective quality assessment results of the human observers and find that the proposed method achieves effective performance in big data environment to evaluate human motion quality.     

Real-time optimal control of automatic sewing considering fabric geometric shapes

This study aims to improve the sewing accuracy and robustness of a control system for automating sewing operations. In a previous study, we experimentally demonstrated a control system for automating the butt seaming of two fabric pieces (one curved and another straight) into a 3D shape. However, because the control method was obtained through trial and error, there were problems, such as low sewing accuracy and low robustness, including sewing failure when changing the sewing machine feed rate. Therefore, this study improved the sewing accuracy using a real-time optimal control. Furthermore, robustness was enhanced using the feed rate and geometric shapes of the fabrics. The experimental results confirm that the proposed method improved sewing accuracy and robustness.     

基于运动控制卡的激光喷丸数控系统设计

为了实现对激光器和工作台的一体化控制,提供操作上的便利,提出了一种基于运动控制卡的开放式激光喷丸强化数控系统的控制方案,PC机完成喷丸强化控制系统的核心处理部分,运动控制卡和单片机完成控制的细节,阐述了硬件结构原理以及软件模块的划分和实现。结果表明,该系统操作方便、开放,便于扩展。     

基于最优控制理论的配网自动化应用研究

在电力系统中应用最优控制理论可以对各种方案进行优化,从而使得电力系统的运行实现自动化、智能化发展.本文从最优控制理论出发对配网自动化的相关问题进行分析和探讨.     

Stochastic LQ-optimal control for 2-D systems

Stochastic LQ-optimal control problem is solved for linear 2-D systems via 2-D polynomial techniques. All stabilizing controllers are parameterized and then minimizing one is obtained. Its transfer function results from two 2-D spectral factorizations and a couple of linear equations in formal power series. A practically useful sub-optimal controller is also considered that results from truncated spectral factorizations and a couple of 2-D polynomial equations.     

基于全相位IDCT滤波器的内插重采样分层编码技术

该文提出了一种基于全相位逆离散余弦变换(APIDCT)内插滤波器的重采样分层编码技术。全相位数字滤波器(APDF)是一种新型的线性相位滤波器。该文给出了APIDCT的设计方法,并通过比较几种APDF的性能,证明APIDCT在图像内插方面性能优异。在此内插方法基础上,结合重采样和自适应算术编码形成了分层图像编码技术,这种技术算法简单,适用于空间可分级编码。实验结果表明仅用其中3层,重建图像的峰值信噪比和压缩率就可达到略优于JPEG的效果,并且压缩率高时也不会出现块效应。     

HOSIDF-based feedforward friction compensation in low-velocity motion control systems

The paper describes an application of a recently introduced methodology for modeling of a class of nonlinear systems — Higher-Order Sinusoidal Input Describing Function technique (HOSIDF) — to a motion control platform for which a precisely controlled motion at low velocity is required. One of the key challenges for these systems is to compensate for the friction, which is particularly difficult to model at low velocities. The frequency-domain HOSIDF modeling framework is used to assist in designing a feedforward compensator. Experiments with a laboratory benchmark system (gimballed camera platform) prove the technique useful.     

基于状态方程的二级倒立摆最优控制研究

主要利用状态空间方程,研究二级倒立摆的最优控制理论。通过对系统进行数学建模,再在平衡点处进行线性化处理、近似计算,得出控制矩阵;接着设计最优控制方案和控制器,以消耗能量最少为标准;最后根据计算结果,通过Matlab程序仿真,得出了上下摆的角度和小车位移的曲线图,并得出结论说明控制器的设计是成功的。     

Sensitivity in optimal control systems

Some recent developments in the area of sensitivity in optimal control systems are summarized. The several sensitivity definitions, analysis methods, and design techniques are placed in perspective. A selected list of references is included.     

拥塞高速数据通信网优化控制模型

针对多信源——信宿高速数据通信网，引入N用户非零和微分对策优化控制路由模型。通信网中每个用户(决策者)，通过决策瞬时队列、路由及通信流量速率，使优化指标(损耗函数)达到最小；给出了反馈Nash平衡解的充要条件，并进行讨论。     

人脸外轮廓线的提取方法

传统的边缘提取算子提取人脸的外轮廓线很难达到理想的效果。文中提出一种基于人脸的几何特征的外轮廓线提取方法：首先采用结合SOBEL和CANNY算子提取边缘信息,然后剂用人脸外轮廓特点去掉杂散的冗余边缘和边缘修补。该算法能够较好地提取人脸的轮廓线,而且定位精度高,满足人脸轮廓特征提取的要求。采用本方法对50多幅人脸图像进行实验,结果验证了方法的有效性。     

Accurate contour control of mechatronic servo systems usingGaussian networks

This paper presents a method of contour control of mechatronic servo systems by using neural networks. The neural network learns the inverse dynamics of the mechatronic servo system. The input data for the mechatronic servo systems are modified from objective trajectories by using the neural network. The Gaussian network is adopted to construct the inverse dynamics of the mechatronic servo system because the Gaussian function is well defined, and its structure and initial parameters can be systematically selected such that the initial network approximates the inverse dynamics of the mechatronic servo system. The actual input/output data of the mechatronic servo system are used for the learning of the Gaussian network. Effectiveness of the proposed method is assured by experimental results of contour control of an X-Y table     

基于SUSAN算子的嘴部轮廓特征提取方法

针对人脸识别中的嘴部定位部分轮廓缺失问题,提出了一种由粗到精的快速准确的嘴部轮廓特征提取方法。该方法是在人脸检测的基础上采用改进的SUSAN算子提取嘴部边缘轮廓信息,通过几何特征和搜索区域相结合的方法实现了嘴唇轮廓的精确定位。仿真结果表明,该定位方法快速有效,正确率较高。     

双线性系统下的量子系统最优控制

借用宏观领域中双线性系统的研究成果,结合量子系统数学模型所具有的双线性系统形式的特点,通过对含有两点边值条件的最优控制器的推导,对量子力学系统进行最优迭代控制器的设计,并通过仿真实验对设计的量子控制系统的状态转移控制进行系统仿真实验,分别对控制参数的不同取值对控制结果的影响进行了详细的分析.