【作者】 张世辉；

【导师】 孔令富；

【作者基本信息】 燕山大学 ， 电路与系统， 2005， 博士

【摘要】 作为串联机器人的对偶机构,并联机器人具有刚度大、精度高、位置误差不累积等特点,已成为机器人领域的研究热点。目前,并联机器人在航空、航天、海底作业、制造领域、辅助医疗和微机电系统等方面有着广泛而重要的应用。 虽然在并联机器人的实际应用和理论研究中均已取得了大量的成果,但在并联机器人的机构学、运动学、动力学、运动控制、路径规划及实际应用方面仍存在一些挑战性的问题。为了充分发挥并联机器人的优点并扩大其应用范围,本文在河北省自然科学基金的资助下将并联机器人应用于汉字雕刻领域。为了使并联机器人实现汉字雕刻功能,作者对该过程中涉及到的汉字图像处理、雕刻刀路的规划与生成、运动学分析、动力学分析、雕刻力的计算及控制等理论和技术问题进行了深入的研究。 研究了将汉字图像信息转化为并联机器人雕刻刀路信息的通用方法,主要包括:归一化、边缘轮廓提取等汉字图像的预处理算法;基于改进的模板匹配的汉字图像细化算法;基于距离的汉字笔画抽取算法;平面凸、凹雕刻刀路的生成算法;平面刀路到球面等其它复杂曲面刀路的映射算法;刀具姿态实时规划方法等。在此基础上组建开发了包括软硬件在内的并联机器人汉字雕刻系统。 对并联雕刻机器人的运动学和动力学进行了分析和研究。求解了雕刻机器人的一、二阶影响系数矩阵,并基于影响系数矩阵对其角速度、线速度、角加速度、线加速度、力与力矩、惯性力等运动学和动力学性能指标进行了分析,根据所得的性能图谱可选择出各项性能指标都相对较好的机构尺寸。给出了基于自构型快速 BP网络的并联机器人位置正解通用方法并以雕刻机器人为例进行了分析。基于牛顿-欧拉方法建立了雕刻机器人的动力学模型,并利用该模型进行了动力学实验分析。 对雕刻力及其控制技术进行了研究。对影响雕刻力的因素进行了分析,给出了雕刻力的计算方法,特别研究了常用的宋体、黑体、隶书和楷体四种不同字体对雕刻力的影响。提出了基于模糊神经网络的雕刻力控制方法,并与传统的阻抗控制技术进行了比较,实验结果验证了该方法的有效性和可行性。 设计了基于 B/S 模式的并联机器人远程汉字雕刻系统的总体结构方案。对其中的视频信息实时传输、远程雕刻信息的提交和解析、任务控制等技术进行了研究和实现。更多还原

【Abstract】 As the counterpart mechanism of serial robot, parallel robot has merits such as high mechanical stiffness, high precision and no accumulation of position errors etc, and has become a hot research topic in domain of robot. At present, parallel robot has found its wide and important applications in such areas as aviation, astronautics, submarine engineering, manufacturing, computer aided medical equipment and mirco eletro-mechanical system(MEMS) etc. Nevertheless, there still exist some challenging problems in mechanics, kinematics, dynamics, motion control, path planning and practical applications of parallel robot in spite that a large amount of achievements have been obtained in theory investigation and practical application. Supported by natural science fund of Hebei province, parallel robot will be applied into area of engraving Chinese characters in this dissertation to fully exert its merits and extend its application area. To make parallel robot engrave Chinese characters successfully, the author implements deep research on related theory and technology such as image processing of Chinese characters, planning and generating of tool path for engraving, analysis of kinematics and dynamics, calculation and control of engraving force etc. A general method for transforming image information of Chinese characters into engraving tool path information of parallel robot is studied. It mainly includes the following contents:algorithms for image preprocessing of Chinese characters such as normalization and contour extraction etc; an improved thinning algorithm of Chinese characters’ image based on template matching and a novel stroke extraction algorithm based on distance; protruding and concave tool path generating algorithm for engraving on plane; tool path mapping algorithm from plane to sphere and other complicated curved surface; real-time planning method of cutting tool’s guise etc. Chinese characters engraving system based on parallel robot, including both hardware and software, is organized and developed also. Kinematics and dynamics of parallel engraving robot are studied. The first and the second order influence coefficient matrices are computed, and mechanical performances for kinematics and dynamics such as angular velocity, linear velocity, angular acceleration, linear acceleration, force and moment, inert force are analysed based on its influence coefficient matrices. So the mechanism size with better performance for every index can be selected according to the performance atlas. A general forward displacement solution for parallel robot based on self-configuration quick BP neural network is presented, and analysis result for the engraving robot is given. The rigid body dynamics model of engraving robot is established based on Newton-Euler method, and dynamics experiment is also analysed with this model. Engraving force and its control technology is studied, the factors influence on engraving force are analysed, and calculating method of engraving force is given. Especially, the influence of different fonts such as Song, Hei, Lishu and Kai on engraving force is studied. The control method of engraving force based on fuzzy neural network is proposed, and its difference between traditional impedance is also discussed, experiment results verify the validity and feasibility of proposed method. The overall system scheme of remote Chinese characters’ engraving for parallel robot based on B/S mode is designed. The several technologies such as video information real-time transmission, submission and analysis of remote engraving information and task control etc are deeply studied and realized.更多还原