【作者】 郭源博；

【导师】 张晓华；

【作者基本信息】 哈尔滨工业大学 ， 电气工程， 2007， 硕士

【摘要】 视觉伺服技术是利用计算机视觉原理,对采集到的图像信息进行快速处理和特征提取,进而提供反馈信息实现系统的闭环控制。视觉伺服系统是以视觉作为获取外界信息的途径,具有信息量大和智能化水平高等特点,目前在工业机器人、微操作机器人和空间飞行器对接系统等领域得到了广泛应用。吊车作为一种搬运工具,在工业生产中发挥着重要作用,如何消除货物在吊运过程中的摆动,进而提高吊车的工作效率一直是控制领域研究的经典问题。抑制摆动的基础是实现重物摆角的有效测量,传统的摆角测量装置安装于吊车系统内部,是一种接触式的测量方法,具有机械结构复杂、独立性差等缺点。本文将视觉伺服技术引入到吊车防摆控制中来,利用计算机视觉的方法实现重物摆角的无接触式测量,并采用逆系统非线性控制策略实现吊车系统的定位防摆控制。本文首先介绍了摆角视觉测量系统的硬件组成,针对摆角的计算机视觉测量需求,本文提出了一套快速有效的图像处理和特征提取方法,并对这种测量方法的可行性进行了分析与实验。结果表明这种计算机视觉测量方法具有较高的分辨率和较宽的测量范围,能准确地测量出钢丝绳的摆角,并满足控制的实时性要求,可以用于吊车系统的定位防摆控制中。根据吊车系统的物理模型,本文应用分析力学的拉格朗日方程建立了吊车系统的数学模型。从本质上看,吊车是一个多变量、非线性、强耦合的欠驱动机械系统,它的控制输入个数少于系统自由度个数,存在着模型复杂、难于控制等特点。为此本文应用逆系统非线性控制方法对吊车系统进行了解耦,进而分别设计位置、绳长伺服控制器和摆动抑制器,提出了变绳长情况下吊车定位防摆控制策略。仿真实验结果表明,该控制方案能够有效地实现吊车系统的准确定位和防摆。最后在吊车实物实验系统平台上,采用摆角计算机视觉测量方法,应用逆系统非线性控制策略,进行了实物实验,实现了基于视觉伺服的吊车定位防摆控制,同时所设计的控制器对于重物质量的变化具有很强的鲁棒性。更多还原

【Abstract】 Vision servo technique utilizes the computer vision theory for high-speed image information processing and feature extraction to realize closed-loop control of the system based on the feedback information. Using vision as its main approach to obtain outside information, vision servo system features extensive information and high intelligence and therefore is presently widely used in the field of industrial robot, micro-operation robot and spacecraft docking system. At the same time, crane, as a kind of transfer tools, plays a significant role in industrial manufacturing. Consequently, how the swaying during goods lifting can be eliminated to achieve high working efficiency of the crane has always been a classical concern for control study. The key point to damping out swaying is the accurate measurement of the angle. Using contact measuring method, traditional angle measuring devices, however, are installed inside the crane system and characterize complicated mechanical structure and poor ability of independence accordingly. The thesis introduces vision servo technique into the anti-swaying control over the crane. Moreover, non-contact measurement is achieved based on computer vision method and inverse system nonlinear control strategy is employed to realize the positioning and anti-swaying control of the crane system.This paper initially focuses on hardware components of angle vision measuring system. In the light of computer vision measurement requirement of the angle, a set of rapid and effective image processing and feature extraction methods is proposed and the feasibility of these methods is analyzed and tested as well in this paper. The experimental results show that the computer vision measuring method, which has a big distinguishable rate, a wide range of measurement, can guarantee the measuring accuracy of the angle and satisfy the need of real-time control; therefore can be used in the positioning and anti-swaying control of the crane system.Based on the physical model of the crane system, this paper builds up a mathematical model in terms of Langrange equation of analytical mechanics. In essence, as a multi-variable, nonlinear, strong-coupled underactuated mechanical system, the crane is complex in model and difficult to control because its number of control input is less than that of system’s degree of freedom. To cope with this problem, the inverse system nonlinear control method is applied firstly for the decoupling of the crane system, then a position and cord length servo controller and swaying controller are designed to put forward the positioning and anti-swaying control strategy of the crane in case of changeable cord lengths. The simulation results demonstrate that the proposed control strategy can realize the accurate positioning and anti-swaying of the crane system.Finally, computer vision measuring method of the angle and inverse system nonlinear control strategy are employed to conduct object experiments on the crane experimental system platform, and positioning and anti-swaying control over the crane is realized based on vision servo thereby. Moreover, the designed controller has good robustness for variable mass of the payload.更多还原