【作者】 陈丹；

【导师】 方康玲；

【作者基本信息】 武汉科技大学 ， 控制理论与控制工程， 2007， 硕士

【摘要】 随着科技的进步,人们对机器人的要求也越来越高,不仅要求机器人能够实现常规的PTP运动(点对点运动)和简单的CP运动(如直线运动和圆弧运动),而且还要能够沿着特定路径进行高速运动,实现运动过程中时间最短。为此,本文对沿着特定路径运动的机器人的时间最优轨迹规划问题作了深入研究,成功运用了三次B样条对由一系列离散路径点组成的机器人特定路径进行了逼近,并利用遗传算法,沿着逼近所得到的机器人路径,对机器人进行了时间最优轨迹规划,从而保证机器人在不偏离原有路径的基础上,实现时间最优运动.首先,本文对机器人应用进行了介绍,全面地总结了国内外有关机器人时间最优轨迹规划方面的研究成果,针对目前机器人轨迹优化中存在的不足,依据时间最优轨迹的原理,提出了基于遗传算法优化的B样条曲线。接着,分析了机器人轨迹规划的一般性问题,在关节空间的轨迹规划中,针对不同的条件分别详述了三次多项式函数插值、高阶多项式插值和抛物线连接的线性函数插值方法;在笛卡尔空间中,则深入讨论了线性函数插值和圆弧插值。然后,本文在研究了机器人时间最优轨迹规划的基本原理,遗传算法的优化原理及B样条曲线的计算方法的基础上,采用B样条曲线构造机器人的轨迹,并利用B样条曲线的分段处理性特点,提出了一种基于遗传优化算法的机器人时间最优轨迹设计方法。在该方法中,机器人沿整个B样条曲线运动的总时间的优化,可转化为对机器人沿每一段B样条曲线运动所需的时间的分段优化,优化时考虑了包括关节角速度和角加速度2种约束条件,并在关节空间对KLD-600机器人的三个关节(腰、臂、腕)进行了优化计算。通过Matlab软件,对轨迹规划算法进行仿真,实验结果表明机器人各关节达到节点总的时间明显减少,各个关节的轨迹曲线平滑,从而验证了该算法的有效性。最后,简单介绍了KLD-600机器人的硬件组成,研究了PMAC运动控制器与主机的通讯原理,利用VC创建了人机对话框,通过调用Pcomm32动态链接库实现上位机与下位机的实时通信,实现机器人的示教操作,并根据D-H方法给出了KLD-600机器人的正逆运动学模型的具体求解过程;通过对机器人示教操作中所得到的路径点进行优化计算,得到最短时间轨迹,从而实现了机器人时间最优的轨迹规划。更多还原

【Abstract】 As the science and technology growing, the demand of the robot’s performance become higher, robots should not only be capable of moving in conventional PTP manner and simple CP manner, but also be capable of moving along a specified path of arbitrary shape at high velocity, in order to realize the less time in the motion. To satisfy such demand, a profound research about the time optimum trajectory planning of robots along a specified path is carried out in this paper. The specified path is assumed to be composed of a series of discrete path points, and the whole time optimum trajectory planning is decomposed into two steps: in the beginning, cube-spline is employed to approach these discrete path points, then genetic algorithm is employed to generate a time optimum trajectory along the B-spline.With the algorithm, the robot is guaranteed to moved in an optimum manner and not deviate from its original path.Firstly, the current situation and development prospect of the technology and applications for the robot was introduced in this paper. The research achievements of time-optimal trajectory planning of robot at home and abroad were summarized. At the same time, the drawbacks of the current trajectory planning were present. According to principle of time-optimal trajectory planning, the optimization of B-spline based on genetic algorithm was proposed. The optimal B-spline was applied in the time-optimal trajectory planning of robot. In this way, the work efficiency was significantly improved, and then general problems in robot trajectory planning were analyzed. In joint space, cubic interpolation, high-order polynomial interpolation and linear interpolation which was joined by the parabola were detailed accounted separately at different situations. In Cartesian space, linear interpolation and arc interpolation are discussed deeply.Secondly, the basic principle of robotic time-optimal trajectory planning was studied. The paper discussed the theory of genetic algorithm and the calculation method of B-spline for constructing trajectory of robot. A method of time-optimal trajectory planning based on genetic algorithm was proposed because of the property of B-spline.In this way, the total time which used by the robot sporting along the whole B-spline could be translated into the sum of part time of the robot sporting along each part of B-spline. The constraint conditions of joint angle velocity and joint angle acceleration were discussed. The wrist joint, manipulator joint, and waist joint of KLD-600 robot were optimally calculated. The algorithm of trajectory planning simulations implemented by the Matlab, its result showed that the total time of each joint arriving the target point reduced, and the trajectory of every joint is smooth, so it demonstrated the efficiency of the proposed algorithm.Finally, the hardware construction of KLD-600 robot was introduced. The communication principle between PMAC motion controller and master was studied. The dialog box of user and machine was built with VC. The real-time communication between main-machine and submitted-machine was realized by calling the Pcomm32. The robot kinematics model and inverse kinematics model of KLD-600 were set up through D-H method. The optimal path points in Cartesian space in teaching operation were obtained by genetic algorithm, in this way, the time-optimal trajectory planning was implemented.更多还原